Talk:Earthquake prediction/Archive 7

Archive 1

←

Archive 5

Archive 6

Archive 7

Archive 8

Archive 9

Archive 10

Proposal

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

I'd been waiting for the RFC to close before making any edits; and also, the more reading I do, the more I realize that I still need to do more reading. However, based on Elriana's encouragement, I've prepared a proposal in my sandbox, for a revision to the sections on VAN method. I've included a little bit of material on the situation post-2001, and my other main goal was to cut down on the length of the sections. Also, I've pulled some material from old versions of the article into my sandbox -- I haven't formed any opinions yet as to whether any of this material has any merit.

Your comments are invited; or feel free to just start editing this page -- I'll revert anything I feel is not NPOV.

https://en.wikipedia.org/wiki/User:JerryRussell/sandbox

JerryRussell (talk) 22:55, 31 July 2016 (UTC)

The most difficult part is to simplify, but it is a necessity. You have to put everything in your sandbox to have all info and see it as a whole, and then start cutting down, day by day at least, leaving enough time to forget and see it fresh again. It is a process that will take more than a week, and my opinion is that you should be left to do it without us commenting in between. When you need an answer you can ask for opinion and when you think you have finished simplification, we can comment on wording. Thank you for your time and effort. --IP202-46.198.213.62 (talk) 09:39, 1 August 2016 (UTC)

Jerry, I strongly suggest that you drop your "History of research programs" section. That looks to be mainly what you have recycled from older (pre-2013) versions of the article, and it is mostly out-of-date. Indeed, much of that is from earliest version of the article (circa 2004), and reflects what was most eagerly anticipated at that time. While I could see an argument for a short "history of earthquake prediction" section, I don't see that all of the EP history since 2004 (relative to the whole) would amount to more than two sentences. I would suggest great wariness in taking any pre-2013 material as it was mostly junk.

BTW, I took the liberty of adding a "Notes" section with a {{reflist}} for you. I hope that is okay. You should also put the citations in their own section. ~ J. Johnson (JJ) (talk) 22:35, 1 August 2016 (UTC)

JJ, the only materials I'm seriously proposing for your consideration at this point would be the first two sections, which I'm using as a testbed for possible updates to the existing sections on the VAN method.

The remainder of the materials in my sandbox are indeed pulled from ancient versions of the page, and I doubt very much that they'd be useful without major updates or condensation, if useful at all. My apologies if I wasn't clear about that. I wouldn't be surprised if much or all of it is junk, as you say. My major purpose in pulling this materials from the archive, was for my own education on the process that has led the article to its current state.

I agree that the article is greatly improved overall from early versions.

IP202's advice is good, to keep working off-line on the material in the sandbox. But rather than waiting for that process to complete, and then introducing all the changes from my sandbox at once, I'm going to make just a few updates, and see if we can get a consensus on removing the tags. JerryRussell (talk) 22:54, 1 August 2016 (UTC)

Done with edits for today. Tags are gone, I hope they don't come back :) Citations to new references are not in CITESHORT format and do not appear in the 'Sources' section yet. If my changes survive for a day or two without being reverted, I'll learn how to fix this. JerryRussell (talk) 23:43, 1 August 2016 (UTC)

Thanks for taking the tags out. I would make two suggestions: it is much clearer to see where you have edited if you do so at the section level (not by hitting the [Edit] tab at the top of the article). And it really is better – and even easier, in the end – if the full citations go straight into the "Sources" section.

I object to any mention of the alleged 2008 prediction, but no more time for that today. Probably should have its own subsection anyway. ~ J. Johnson (JJ) (talk) 00:31, 2 August 2016 (UTC)

@JerryRussell, thank you for braving the criticism and attempting to move forward with this. Comments:

1)It could be mentioned that 'natural time' is a concept developed for critical phenomena. That would at least give some hint as to where to look for more background on the term. I think this should be an acceptable addition to both sides of the debate on how to describe VAN.

2) I don't object to the inclusion of the 2008 event, so long as it is clear that there is serious doubt as to the validity of the prediction (how it is presented now is ok). However, the coverage of this event is one of the most contentious issues on this talk page. If that conflict persists, I would propose separating this 'prediction' from the rest of the VAN predictions subsection. Get consensus on the pre-2001 section, and then work on the separate section on the 2008 stuff. I think this is acceptable because of the reformulation of VAN techniques around 2001 to include 'natural time' and initiation of regular status reports to the Cornell library archive in 2006. These 2 changes should make the 2008 events sufficiently distinct from the 'predictions' that went before to justify a separate subsection, especially if that aids in current consensus and future curation. Elriana (talk) 01:03, 2 August 2016 (UTC)

JJ and Elriana, thanks for your comments. IP202 and I have taken turns at tweaking the section, hopefully it's all for the better.

I'm not sure I agree that the 2008 prediction deserves to be in its own subsection: wouldn't this only call more attention to it? As it stands, there's only one short paragraph about this material, which would make for a very short stand-alone subsection. VAN advocates claim to have made a new breakthrough with 'natural time', and to have cleaned up their act by posting public predictions, but the detractors say their behavior is just more of the same. I tend to agree with the latter appraisal. But if either of you made the edit to create a new subsection, I wouldn't revert it. JerryRussell (talk) 19:07, 2 August 2016 (UTC)

I only suggest making the 2008 stuff its own subsection if that serves to narrow the scope of any editing conflicts. Otherwise, I generally agree with @JerryRussell. Elriana (talk) 20:44, 2 August 2016 (UTC)

Jerry: per WP:BRD I have reverted your bold revision of 23:33 yesterday where you changed the subsection title from "1983–1995: Greece (VAN)" to "1983–2008: Greece (VAN)", extending the period covered. I am adamantly opposed to this extension.

The period covered (to 1995) was chosen for very good reason: 1996 marked two comprehensive and in-depth reviews of VAN (the Lighthouse debate and the GRL debate) to 1995. There has been no similar comprehensive reviews after that, from which IP202 (repeatedly disparaging all of the critical papers from 1996 as "two decades old publications" that are no longer applicable) then claims that there has been no new criticism, and therefore VAN is accepted science. (Even stating on 27 July that: "Discussion is really over, everything claimed by my point of view is sourced.")

Another factor is that prior to 1995 VAN churned out a lot of predictions, afterwards not so many. (Elriana is correct: post-1996 is distinct.) While the subsequent predictions are not entirely lacking in interest, individually they do not have same level of notability as predictions that have been excluded from this article for reasons of space. Adding the VAN post-1996 predictions, either collectively or individually, gives VAN undue prominence.

More particularly, the extension to 2008 (instead of to, say, 2015) is a special adjustment to get in the 2008 alleged prediction. As we have seen in the discussion at #Rfc: Item d: removal of "natural time" material as being fringe, the reason for getting that in is to have a coatrack for getting in "natural time". This is fundamentally against the policy of WP:NPOV.

Regarding the 2008 alleged prediction: for the reasons above it (nor any post-1996 event) belong in the current section. But (as stated just above) nor does it warrant its own section. It simply is not notable (outside of Greece), and its only significance here is to burnish "natural time". That amounts to promotion of fringe material.

I see that IP202 has been busy spinning various points in favor of VAN. Lack of time to comment now is NOT assent. ~ J. Johnson (JJ) (talk) 23:32, 2 August 2016 (UTC)

Hmm... the 2008 event has been construed by many sources as a 'prediction'. IP202's latest edits look pretty neutral to me, and where they didn't look neutral I tried to fix them. Seems like you have many other quarrels with the latest edits? Maybe you should put the POV warning tag back up? I'm not going to: changing the date in the title to 1995, without changing the contents of the paragraph itself, just looks like a typo. JerryRussell (talk) 01:08, 3 August 2016 (UTC)

Yes. That paragraph should come out. Whether I tag (yuck) the other questioned material or not, I will (like IP202 doesn't) explain my objections. (But my time is currently rather overloaded.) A note in passing: just because some number of sources say there has been a prediction is rather dubious when 1) those sources tend to be same small group of proponents, and 2) they still can't point to a clear (echoing one of those ancient 1996 criticisms), explicit prediction. As I keep saying: where's the beef? ~ J. Johnson (JJ) (talk) 21:50, 3 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Forecasting vs. prediction

I'm getting the impression here that 'earthquake forecasting' is considered a legitimate seismological study, as opposed to 'earthquake prediction' which is rather disreputable, and seemingly impossible to accomplish on any routine basis that would be socially useful.

This leads me to wonder if this article should be re-named as "Earthquake Forecasting", with "Earthquake prediction" as a re-direct to the new title? The article could then be re-organized, with the highlight on trend methods such as seismicity patterns, etc.

Does anyone know of serious websites where earthquake forecasts are posted for the public? I found that Keilis-Borok and Kossobokov have a site where current M8/Msc forecasts are posted, but it's password protected. And I found a really cheesy-looking anonymous website at www.world-earthquakes.com with many intriguing results posted. Surely there must be something better, but I couldn't find it.

I will say this: if forecasting methods were saying that the big one was coming to the Pacific Northwest, and then radon readings went sky high, geo-electromagnetic precursors appeared at all wavelengths, and the cows headed for the hills, I wouldn't consider it superstitious to get out of Dodge if I could. I could use a vacation anyhow.JerryRussell (talk) 19:30, 2 August 2016 (UTC)

Oh and by the way, I would not want the said information to be broadcast on the local news. A panic would cause crowds at the airport, and make it harder to get a ticket for my getaway. JerryRussell (talk) 20:32, 2 August 2016 (UTC)

Earthquake 'forecasting' is indeed what most seismologists will call their long-term, history- and geologic map-derived statistical predictions. However, most lay-people do not know this distinction. And, as demonstrated by the many precursors scientists have studied over the years, short term predictions are very much something we would like to be able to do. Just because we haven't managed to make successful predictions doesn't mean we haven't tried. I think it is very important for this article to convey the history of the most notable efforts. Explaining that these efforts have been made and discussing why they failed demonstrates why the longer-term forecasts are the best warning scientists have to offer right now. As for the title of this article, that depends on whether you think short-term predictions are a type (and therefore subset) of forecasts, or the long-term forecasts are a type of prediction. I *think* forecasts in any field are normally considered a type of prediction (supporting the current title). But I am open to other perspectives on this.

As for websites with posted forecasts, the best one for the US is http://earthquake.usgs.gov/. The trick is that the information is presented in the form of hazard maps (e.g., http://earthquake.usgs.gov/hazards/products/conterminous/2014/2014_1Hz10pct.pdf), which are based on probabilistic hazard curves for a grid of locations. Parsing the details of these maps can be a bit tricky, as there are multiple ways of turning a grid of curves into a map. Reading the map captions and documentation is essential. If you explore a bit, they have a widget for making a custom 'map of probability of earthquake larger than given magnitude within selected distance' for a user-specified region.

For forecasts outside of the US, the USGS has some (http://earthquake.usgs.gov/hazards/products/foreign/), but a better source for global coverage (slightly out of date, but better than nothing) is the Global Seismic Hazard Assessment Program (http://www.seismo.ethz.ch/static/GSHAP/). Japan's government has made interactive maps available at http://www.j-shis.bosai.go.jp/en/. Elriana (talk) 21:24, 2 August 2016 (UTC)

Elriana, I'm shocked! Don't you know about UCERF3, the Uniform California Earthquake Rupture Forecast? Check it out. You, too, Jerry. It is like the most comprehensive EQ forecast ever attempted. And a very impressive piece of work. (Perhaps we'll have a quiz in the morning. :-)

As to earthquake forecasting, I quote from the lead of this article: "the probabilistic assessment of general earthquake hazard..." Prediction is generally considered (popularly and scientifically) to be the foretelling of the specific parameters of the next "big one". Prediction tends to be about whether to spend the next few nights sleeping out in the car. Forecasting is more about "your URM building is going to come crashing down. Eventually. How much lead time do you want?" ~ J. Johnson (JJ) (talk) 23:54, 2 August 2016 (UTC)

I'll take as much lead time as I can get :) Some buildings only have a useful lifetime a few years out anyhow, and buying earthquake insurance is often an available option. Actuaries need to know how to price those policies. JerryRussell (talk) 01:10, 3 August 2016 (UTC)

I have found UCERF3 before, but I was attempting to find products with more global coverage. (I've had issues before with the US-centric nature of many wiki-articles). The state of the science in areas like Japan and California is much more detailed than that in more remote, less populated and/or less wealthy places. But that is beside the point.

My question: Do we need to devote a little more text to the statistical and historical risk analyses that we have categorized as 'forecast' methods? The term 'earthquake forecast' redirects to this article rather than having one of its own. Should there therefore be a section that (briefly) describes the state of these forecasts and directs people to some of the relevant wikiarticles on specific hazard assessments? Comments and ideas from all welcome. Elriana (talk) 18:54, 15 August 2016 (UTC)

@Elriana: Oops, I didn't this before. I hope you didn't think I was ignoring you!

I think earthquake forecasting should have its own article, as these are both large subjects with distinct differences. That that article is not yet written – well, would you be interested? Even if it is little more than a stub, that would be better than trying to temporarily squeezing it into this article.

UCERF3 stands alone. In one of the reports Field et al. note that very few other regions have the kind of voluminous, detailed data necessary for this kind of work, and even in California, which is arguably the most seismically studied region in the world, they found the data barely adequate. ~ J. Johnson (JJ) (talk) 23:21, 25 August 2016 (UTC)

I'll contemplate what should go into an Earthquake forecasting article. I probably have a fair number of relevant references (and textbooks) already, but time is not my friend at the moment. I'm not sure how one goes about un-redirecting one term from another, but I'm sure I can find instructions.

I would note, also, that if we can't nail down the statistical distribution of earthquake occurrence in most of time and space because of a severe lack of data, it is really no wonder we have trouble understanding the events well enough to predict them. As is common with earth processes, a little data leads to some great theories, but testing those theories is going to require several orders of magnitude more data points. The amount we do not know about these complex systems is why I got into earth science in the first place.Elriana (talk) 22:40, 1 September 2016 (UTC)

Hi Elriana & JJ, I also missed that this thread is ongoing. Sorry!

I guess I don't understand why JJ objects to starting with a new section in this article? Then when we have enough content, we could fork the new article.

I'm not really sure I even understand what the key issue is, defining the difference between prediction and forecasting. Is it a matter of short-term vs. long-term, or probability estimating vs. binary alarms that can then be judged as hits or misses? My reading of the literature is that probability is the key issue when this distinction is made, but I could be wrong. I think a lot of the precursor methods described (maybe all of them) will really never be anything more than short-term forecasting methods, rather than prediction methods. That is, they should be interpreted as probability estimates. This is why I suggested that maybe this entire article should be re-named as 'forecasting'. Another potential problem is that we can certainly come up with a clear definition of the difference between 'forecasting' and 'prediction', and indeed there might be a high level of support for those definitions in the literature. But, we're far from a situation where all the literature is in agreement. I think there are some papers that treat 'prediction' as synonymous with 'short-term', and some groups that describe 'forecasts' as 'predictions'. JerryRussell (talk) 23:35, 1 September 2016 (UTC)

Elriana: yes! do look into that. I have a feeling you would do a great job, and I will be pleased to help you on any of the niggling details. Starting with your textbooks is probably the best idea. I have some references I'll send to you (including some reviews), with a special recommendation to read Zechar 2010. Also take a look at Field et al. re UCERF3 to get a feel for what real forecasting requires.

When you think you have a reasonable draft there are several ways to proceed, but I would recommend starting in the "Draft" name space. (Just go to Draft:Earthquake Forecasting, and take the offer to create that page.)

Jerry: of course not all literature is in agreement, and some authors use "prediction" and "forecast" synonymously. However, what these two terms cover is varied enough that to not follow the lead of most experts in distinguishing these terms at the outset leads to all kinds of muddling, such as you just referenced. And the key difference shouldn't be so hard: prediction is particular about time, location, and magnitude, while forecasting is a probablistic estimate over some range of time (usually decades), location, and magnitude. Is this not made clear enough in the lead?

Yes, the boundaries are not sharp. And again yes that prediction tends to be short-term. That is because near-term phenomena are clearer. E.g., it is easier to predict whether Florida will be hit by a hurricane in a particular week if the week in mind is next week, not the same week a year out. But note that some alleged bases for prediction are cyclical (e.g., tidal forcing), and so, theoretically, can "signicantly contract" the time window many years out.

Another distinction is whether an alarm is issued. E.g., Portland has large numbers of URM buildings. If we had definite and certain knowledge that The Big One was going to hit tonight than an appropriate response is, indeed, to sleep in the car for a night or two. The appropriate response to a forecast that there is a certain probability of an M 6.7 earthquake in the next 50 years is to reinforce or tear-down those buildings. Which won't do us a damn bit of good if The Big One strikes tonight. But in the long run of things it would be useful to start reinforcing those buildings.

I object to including earthquake forecasting in this article because the terms are different, and this article is big enough, When Elriana has some material together it should go straight into the new article (either main space or Draft space). ~ J. Johnson (JJ) (talk) 22:32, 3 September 2016 (UTC)

And the key difference shouldn't be so hard: prediction is particular about time, location, and magnitude, while forecasting is a probablistic estimate over some range of time (usually decades), location, and magnitude. Is this not made clear enough in the lead? I'm sorry, I'm not finding this clear at all. What if someone says "a 50% chance of an M8+ EQ in the next two years somewhere in the Cascadia Subduction Zone", is that a forecast or a prediction? "Magnitude 6 (+/- .7) somewhere in Greece in the next 22 days", forecast or prediction? I can see why someone might be hesitant to call these 'predictions', but they don't exactly fit your definition of 'forecast' either. JerryRussell (talk) 23:25, 3 September 2016 (UTC)

A key concept is probabilistic estimate. So this "50% chance" in some extended period is a forecast. As such it is not confirmed or disconfirmed by a single event, or lack of same, but whether, in the long run of a sequence of such forecasts, actual events happened at that rate.

Your second example is, of course, the VAN predictions. Note that the objection that VAN's utterances are not truly predictions does not mean they are forecasts; VAN just isn't specific enough that their utterances can be subjected to any kind of scientific test. Personally, I think that "predictions" with ranges as large as a month and all of Greece are better handled probabilistically; that is, as forecasts. But VAN insist they are doing predictions (originally on the order of seven hours), and justify themselves on the basis on specific events, so calling their utterances "forecasts" would just confuse everyone.

Another key concept is the scope of particular. As you have noted below (#Prediction and forecasting, two sides of the same coin?), the ICEF says a "prediction involves casting an alarm". (And isn't that just what I said just above: "... whether an alarm is issued"?) As you saw in Nature Debates, Geller sees earthquake prediction as "an alarm of an imminent large earthquake, with enough accuracy and reliability to take measures such as the evacuation of cities." (Or sleeping in one's car.) With a forecast of (say) an M 6.7 or greater quake in the next 30 years we don't issue alarms to evacuate, we issue advisories to do seismic retrofits. Or convince friends and loved ones to move out of those quaint 1920s URM apartment buildings. ~ J. Johnson (JJ) (talk) 23:15, 6 September 2016 (UTC)

seismo.info, again. previously discussed, and dismissed.

"Surely there must be something better, but I couldn't find it" - of course there is: http://seismo.info, with science behind it and the obvious proof. Fascinating. 31.185.124.108 (talk) 16:00, 25 August 2016 (UTC) No, not fascinating at all, nor scientific. We have already disposed of http://seismo.info/ and Dr. O; see #Dr. Omerbashich earthquake prediction, above. And if one couldn't guess from the same old refrain and language, well, geolocation shows the usual source of this crackpottery. ~ J. Johnson (JJ) (talk) 23:18, 25 August 2016 (UTC)

Possible new subsection: VAN 1996-present

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

IP202, could you provide any additional information about VAN predictions extending from 1996 to the present time? I find the Arxiv article titles to be very cryptic, I have no idea whether there might be 'predictions' embedded in some of those articles. Have any VAN predictions other than the one in 2008 been discussed in any reputable popular publications in Greek?

If JJ wants more data and evidence of notability for VAN after 1996, let's see what we can do to give it to him. JerryRussell (talk) 01:43, 3 August 2016 (UTC)

I answer my own question to some degree: the article on VAN method states --

The VAN team claim to have successfully predicted twenty five of the 28 major earthquakes from 2001 through 2010 in the region of latitude N 36° to N 41° and longitude E 19° to E 27° with this new analysis.

References

The reference is to the book 'Natural Time Analysis'. Searching Google Scholar turns up 91 citations of this book, many of which appear to be independent research groups.

Perhaps someone has a copy of the book, and could post the section on predictions & results, starting on p. 326? JerryRussell (talk) 05:17, 3 August 2016 (UTC)

I will try to find it. In the meantime have a look here. --IP202-46.198.213.62 (talk) 06:51, 3 August 2016 (UTC)

If sources for VAN post-1996 predictions cannot be readily found it does suggest the topic is fringe. ~ J. Johnson (JJ) (talk) 23:45, 3 August 2016 (UTC)

It does seem that the 2008 prediction is the only one from VAN after 1996 that has been reviewed in any secondary sources. Even Mary L-V's book (linked by IP202 above) doesn't have any extended discussion, at least judging by the table of contents. I think this VAN claim should then be treated as a primary source, and as such should be used cautiously if at all?

Accordingly, it does seem most appropriate at this point (also per Elriana's advice) to pull the 2008 prediction out into its own paragraph, and tag it as disputed. I had been hoping for more material for the section, but that doesn't seem to be happening. JerryRussell (talk) 03:11, 4 August 2016 (UTC)

Yes, and yes. ~ J. Johnson (JJ) (talk) 23:30, 4 August 2016 (UTC)

Natural Time Analysis: The New View of Time - Precursory Seismic Electric Signals, Earthquakes and other Complex Time Series Page 296:
Table 7.1 All EQs with Ms(ATH) ≥ 6.0 within N41 36 E27 19 since 2001 along with the relevant SES activities. The cases in parentheses refer to EQs for which the expected magnitude (on the basis of the SES amplitude) was Ms(ATH) ≈ 6.0, but the actual magnitude turned out to be somewhat smaller. The last column gives, in each case, the relevant documentation publicized before the mainshock occurrence, when available. The EQs grouped together refer to almost the same epicentral location.

EQ Related SES activities

Date	Epicenter	Magnitude	Station	Date	Publication
D/M/Y	◦N – ◦E	Ms(ATH) – Mw(USGS)		D/M/Y
26/7/2001	39.05–24.35	5.8–6.5	VOL	17/3/2001	Ref. [41] submitted on 25 March 2001
14/8/2003	38.79–20.56	6.4–6.2	PIR	8/8/2003	Ref. [35]
31/1/2005	37.41–20.11	6.2–5.7	PIR	17/10/2004	Ref. [52]
17/10/2005	38.13–26.59	6.0–5.5	MYT	21/3 and 23/3/2005	Ref. [51] submitted on 16 April 2005 and Ref. [43]
(17/10/2005	38.14–26.59	5.9–5.8	MYT	21/3 and 23/3/2005	Ref. [51] submitted on 16 April 2005 and Ref. [43])
20/10/2005	38.15–26.63	6.1–5.9	MYT	21/3 and 23/3/2005	Ref. [51] submitted on 16 April 2005 and Ref. [43]
18/10/2005	37.58–20.86	6.1–5.7	PIR	17/9/2005	Ref. [42] submitted on 22 October 2005 and Ref. [43]
8/1/2006	36.21–23.41	6.9–6.7	PIR	17/9/2005	Ref. [42] submitted on 22 October 2005 and Ref. [43]
(3/4/2006	37.59–20.95	5.3–5.0	PAT	13/2/2006	arXiV:0602603v1 [25 February 2006] and Ref. [50])
(4/4/2006	37.58–20.93	5.7–5.3	PAT	13/2/2006	arXiV:0602603v1 [25 February 2006] and Ref. [50])
(11/4/2006	37.64–20.92	5.7–5.4	PAT	13/2/2006	arXiV:0602603v1 [25 February 2006] and Ref. [50])
(11/4/2006	37.68–20.91	5.9–5.5	PAT	13/2/2006	arXiV:0602603v1 [25 February 2006] and Ref. [50])
(12/4/2006	37.61–20.95	5.9–5.6	PAT	13/2/2006	arXiV:0602603v1 [25 February 2006] and Ref. [50])
25/3/2007	38.34–20.42	6.0–5.7	PAT	8/2/2007	arXiv:0703683v1 [26 March 2007]
(29/6/2007	39.25–20.26	5.7–5.2	PAT	23 and 24/4/2007	arXiv:0703683v5 [15 May 2007])
6/1/2008	37.11–22.78	6.6–6.2	PAT	7/11/2007	arXiv:0711.3766v1 [23 November 2007] and Ref. [21]
14/2/2008	36.50–21.78	6.7–6.9	PIR	14/1/2008 and 21-26/1/2008	arXiv:0711.3766v3 [1 February 2008] and Ref. [21]
14/2/2008	36.22–21.75	6.6–6.5	PIR	14/1/2008 and 21-26/1/2008	arXiv:0711.3766v3 [1 February 2008] and Ref. [21]
20/2/2008	36.18–21.72	6.5–6.2	PIR	14/1/2008 and 21-26/1/2008	arXiv:0711.3766v3 [1 February 2008] and Ref. [21]
8/6/2008	37.98–21.51	7.0–6.4	PIR	29/2-2/3/2008	arXiv:0802.3329v4 [29 May 2008] and Ref. [21]
21/6/2008	36.03–21.83	6.0–5.6	PIR	5/6/2008
14/10/2008	38.85–23.62	6.1–5.2	missed
(13/12/2008	38.72–22.57	5.7–5.2	PAT	9/10/2008	arXiv:0711.3766v5 [7 December 2008])
16/2/2009	37.13–20.78	6.0–5.5	PIR	12/12/2008	arXiv:0707.3074v3 [5 February 2009] and Ref. [47]
3/11/2009	37.39–20.35	6.1–5.8	PIR	24/10/2009	–
(18/1/2010	38.41–21.95	5.7–5.5	PAT	24/10/2009	arXiV:0904.2465v8 and v9[14 and 27 November 2009])
(22/1/2010	38.42–21.97	5.6–5.2	PAT	11/11/2009	arXiV:0904.2465v8 and v9[14 and 27 November 2009])
(9/3/2010	38.87–23.65	5.6–	LAM	27-30/12/2009	arXiV:1003.1383v1 [6 March 2010])

--IP202-46.198.213.62 (talk) 17:27, 4 August 2016 (UTC)

Hello IP202, thanks for posting this table!

I have reverted your latest addition to the section on natural time and the 2008 prediction. What we need are secondary source reviews agreeing that VAN natural time analysis is successful on a regular basis. I agree with you that the 2008 SES record seems to have preceded the earthquake, and that this event was reported widely enough to be notable. But, I'm not convinced that this was anything but a fluke. Adding more information on this page about a method that might not work, would be undue weight for sure.

I spot checked a couple of items in this table from the book.

The paper arXiV:0904.2465 v9 (27 Nov 2009) discusses SES events of Oct. 24 and Nov. 11, 2009. It also mentions two earthquakes, Nov. 16, 2009 and Nov. 19th (Ms=4.3 and 4.4 respecitively), and says “the system is near the critical point.” The table in the book chapter, p. 296, associates these events with mainshocks occurring Jan. 18 and Jan. 22, 2010.

ArXiV:1003.1383v1 mentions SES activity at the Lamia station from Dec. 27 thru Dec. 30, 2009. It says that “almost three weeks later, two strong earthquakes of magnitude Ms(ATH)=5.7 and 5.6 occurred…” These seem to be the same quakes mentioned in the table above for Jan. 18 and 22, 2010. Yet the table on p. 296 associates these SES with an earthquake of March 9, 2010.

I hope you won’t be offended, but I have to agree with JJ here. Where’s the beef? The ArXiv papers are as vague as can be, they never make any statements that could be construed as predictions, and the observations are buried deep in the texts. How many of the VAN papers at ArXiv include such non-predictions? All of them? How many false alarms embedded in all those papers?

Also there seems to be some confusion of cause and effect, as to which SES signal is to be associated with which earthquake.

Here’s what a prediction would look like: VAN would post an article to Arxiv entitled “Earthquake Ms>6 predicted at X location before Y date”, and then the earthquake would happen. Then everyone would know to stay outside of dangerous buildings for that period, and lives would be saved.

Imagining for a moment that VAN operations were extended to the Pacific Northwest, I don’t see anything in these ArXiv articles that would tell me when to go camping in my friend's teepee. Or am I missing something?

JerryRussell (talk) 18:49, 4 August 2016 (UTC)

• Anyone acting on good faith can apply his/her critique and it is welcome.

• In the first example the arXiv doc is published on Nov 27. The earthquakes that are mentioned above have occurred before the publication and are a lot smaller than M 6, so they are out of the question as predictions, they rather are the closing of nearby seismicity to the critical point.

• The same goes to the second example, only that these earthquakes are quite large and can indeed be confused with the impending earthquake from the specific SES activity, "(but see also Refs.81,82)" → "This type of analysis suggests that the critical point has been approached, which reveals that an impending mainshock, if any, is imminent". The earthquake occurred 3 days after the submission.

• In my opinion, if someone dares to announce small earthquakes (less than M 6) and people go out of their houses every now and then, he is definitely going to receive lawsuits. Public announcements in newspapers are rare and only when a large event is expected for sure and immediately. In the example of 2008 the specific location was not revealed to the public, it was only given after the earthquake happened, in order for the newspaper itself to be protected from lawsuits.

• Referring to "SES activity" along with a timestamp, anywhere in any publication, means there is an impending earthquake after the publication, according to VAN. I guess it will be impossible to find any "SES activity" without its accompanying earthquake, if VAN are correct. But I do not think a scientist will dare to take his chances with the law on this, being as open as we would like. Responsibility comes first and I would also try to hide the info very well.

• I partially understand the reason we only refer to 2008 prediction and not the whole series of predictions, so 2001-2011 became 2008 because of referencing we still miss. But the revert got the article back to the version that mistakenly explains the usage of natural time analysis and should be corrected. I only wonder if the understanding of natural time usage is really needed, as it allowed me to "decode" the examples above (I' m pretty proud I did, really).

--IP202-46.198.213.62 (talk) 23:29, 4 August 2016 (UTC)

IP202, this perspective seems very realistic, yet fascinating. I agree that false alarms of earthquakes could lead to lawsuits. Are there examples that can be documented in RS, or do you know if this something that VAN has experienced? We also know that seismologists who reassure the public that they are quite safe from any earthquake, even when radon levels are soaring, can be prosecuted and convicted for manslaughter. But what if someone were to sue Ethnos for their intentional omission of the life-saving information regarding the location of the 2008 quake? If I were an ambulance-chasing lawyer, the idea would occur to me.

You are saying that the reason VAN does not publish clear predictions, is for legal liability reasons? It does seem to be a high-risk occupation. VAN could find themselves in trouble no matter how they proceed, that is unless they can nail the predictions 100%. I wonder if they can really shield themselves from liability by encrypting the predictions as observations? Couldn't somebody still try to sue them for misleading SES info and algorithms?

I don't see how the material from footnotes 81,82 helps to resolve the confusion of which SES event associates with which earthquake. JerryRussell (talk) 00:32, 5 August 2016 (UTC)

Here is a lawsuite against newspapers, after the 2008 earthquake.--IP202-46.198.213.62 (talk) 09:32, 5 August 2016 (UTC)

Also, IP202, with regards to your edits to the page: please read WP:ADVOCACY and try to be an expert and a steward, but not an advocate for your point of view. JerryRussell (talk) 01:03, 5 August 2016 (UTC)

Disconfusing natural time analysis inside the article is not advocacy. The simple two steps of natural time analysis applied in VAN method are important, as understanding it, the vagueness claimed above disappears; it clarifies what the earthquakes presented after the SES activities, are.--IP202-46.198.213.62 (talk) 09:32, 5 August 2016 (UTC)

The article about lawsuit against newspapers is dated March 4, 2008. Google translates the headline as "Complaint against unknown for spreading false news" and the text (at least in translation) gives no hint as to which newspapers were being sued. Chouliaras is said to be "of the group of supporters" of VAN, and he says there has been no earthquake warning from VAN at that time. From the table of SES signals and ArXiv reports above, that seems to be true. So what's left unclear, is why these unnamed newspapers were spreading rumors about earthquake predictions, and where these alleged predictions came from.

IP202, reasonable people can certainly disagree about whether your edit to the 'natural time' section, represents additional undue weight. I didn't intend any disrespect, I'm just asking you to carefully consider in each case whether your edits are consistent with Wiki policies for treatment of fringe topics. I wish Wiki didn't use that term 'fringe', by the way: I feel it's unnecessarily insulting. The policy is really about all minority views, even respected minorities which might be correct in their views. JerryRussell (talk) 18:13, 5 August 2016 (UTC)

Per the WP:Advocacy nutshell: "Wikipedia is not a venue for raising the visibility of an issue or agenda." Giving "natural time" more prominence than it warrants is advocacy. Likewise for boosting VAN. VAN has been rejected by mainstream scholarship. IP202 is obviously too close to VAN to have a neutral viewpoint on any of this. If he truly wants to contribute to Wikipedia he should consider a different content area.

The supposed threat of lawsuits is VAN's excuse for not publishing their predictions. They want to provide their "predictions" to the government, and let the government take the heat for either publishing, or not publishing. This was discussed in one of the GRL23 papers. Making these quantum-like maybe/maybe-not "predictions" vague and somewhat camoflagued gives them a lot of slack to argue what, or even if, these "predictions" were, or were not, after they see the result. It's the basic approach used by charlatans. (And yes, they have been described as such.) Seismologists, taking heed of the advice of statisticians, have worked out how to deal with these kinds of problems. But VAN seem to consider all criticism as not only incorrect and false, but as an attack, and have seldom taken heed of the constructive suggestions that have been made for improving their work. ~ J. Johnson (JJ) (talk) 22:49, 6 August 2016 (UTC)

JJ, you wrote: "Seismologists, taking heed of the advice of statisticians, have worked out how to deal with these kinds of problems." How so? Are there any seismologists at all, who are routinely making and publicizing short-term earthquake predictions of any kind, whether they're worried about lawsuits or not, and whether they're taking heed of statisticians or not? I thought you've been telling us that mainstream seismology believes that prediction is probably impossible in principle, and certainly not possible within the existing state of the art.

I agree that charlatans would issue vague predictions, but IP202 has provided an explanation of why VAN must act this way whether they are charlatans or not. They are publishing their actual observations and calculations, and perhaps some day an independent group will review their data and offer a judgment as to whether their revised methods are working or not.

In an article about earthquake prediction, in which the mainstream dismisses all methods as impossible: what is the appropriate weight to give to the only prediction method which publishes their papers in peer reviewed journals, and is still making the claims that they can do this successfully? I think it's appropriate to give enough space to provide a basic, rudimentary explanation of the method. And I find IP202's participation to be very helpful in getting the story straight. JerryRussell (talk) 23:18, 6 August 2016 (UTC)

At ArXiv, the most recent submission I could find from Varotsos is from Aug. 2015. As to my earlier question as to which VAN papers include new SES, seismic or natural time calculations, the answer seems to be "all of them". New papers or updates came out regularly right up until one year ago, then suddenly stopped.

Varotsos retirement is probably not the explanation: many professors go right on doing their research after becoming emeritus. But perhaps something has changed just recently. JerryRussell (talk) 23:50, 6 August 2016 (UTC)

I didn't say that seismologists are making short-term predictions (nor do I believe any are), only that seismologists (and geologists and even meteorologists) have worked out how to deal with some of the issues of prediction. E.g. (to take an almost trivial problem), predictions can be numbered, "signed, sealed, and delivered" to a public repository, to be opened at a specified time. And on a standard form, approved by the community of experts, with all of the caveats stated, and all of the boxes filled in.

IP202's explanations are not availing, VAN's excuses are specious. I believe they were investigated at one point for causing public alarm, but that was because they were flogging it on the television and newspapers. IP202's "people go out of their houses every now and then" trivializes the matter. If a large earthquake was predicted for a large city the "countermeasures" would amount to a shutdown of business, causing economic losses of billions of dollars.

And it is incorrect to call VAN "the only prediction method which publishes their papers in peer reviewed journals}. If you mean that VAN publishes their papers in peer reviewed journals, well, yeah, but then so do all the other scientific predictors. But if you mean that only they publish their predictions in peer reviewed journal, sorry, wrong. First, they haven't published any predictions, only their a posteriori claims of prediction. Second, Bakun and Lindh did publish their prediction (in Science), and Kellis-Borok published several predictions. (Have you read the article??)

As to the current status of "natural time", check also the current activity of Sarlis and Skordas. ~ J. Johnson (JJ) (talk) 02:55, 7 August 2016 (UTC)

OK, the predictions by Bakun & Lindh and Keilis-Borok didn't use non-seismic precursors, but they were bordering on short-term predictions: Bakun & Lindh gave a 5 year period, and the Keilis-Borok predictions seem to have time frame less than a year. So they're somewhat relevant. I really have read the rest of the article, but perhaps have been focusing too much attention on the VAN aspect, as the problem at hand.

I'll let IP202 respond to your other comments if he cares to; your suggestion to post predictions to a temporarily confidential repository does seem to be a way to avoid the potential lawsuit problem, if VAN wanted to subject their method to objective statistical testing. JerryRussell (talk) 03:54, 7 August 2016 (UTC)

The state of your last "if" seems to be "no", as in they do not want to. Various very reasonable suggestions were made twenty years, which they rejected. Varotsos in particular shows a tendency to accept some reply, technique, or study as being absolutely authoritative, admitting of no further discussion. ~ J. Johnson (JJ) (talk) 23:29, 7 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

"Classification" and a non-NPOV

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

JerryRussell (talk) 18:22, 10 November 2016 (UTC)}}

The recently added section on "Classification" of long, intermediate, and short term prediction is, at best, rather pointless, making arbitrary distinctions of little use. It is also inaccurate. But under that vapid surface is a deep running current of non-neutral POV.

The first clue is the phrase "[e]xcluding solely negative views, its progress has already reached the statistically proven stage". The "solely negative views" to be excluded are, as we have seen in this discussion, those of all those seismologists who say earthquake prediction (in the more limited sense of prediction) is impossible, while the "statistically proven stage" refers to VAN, and the VAN proponents' own estimation of "statistically proven".

Any doubt of this POV is dispelled in the next paragraph with: "Precursors [...] are mainly non-seismic and mainstream seismology instrumentation cannot detect them." This is the VAN party line, as formulated by Uyeda: that seismologists look at only seismic phenomena, and are too incompetent to read up about the latest non-seismic precursor (i.e., "natural time"). This is entirely and utterly false, as can be seen by even cursory examination of the sources for the Earthquake prediction#Precursors section. E.g., the ICEF 2011 report explicitly addresses electromagnetic signals, including VAN SES signals. The search for earthquake precursors has been far-reaching (e.g., see the report of the IASPEI - the International Association for Seismology and Physics of the Earth's Interior), but surely everyone here can figure that out themselves from the sources.

The inherent promotion of VAN in the "Classification" section shows yet again the persistent non-neutral POV of the anonymous editor "IP202". ~ J. Johnson (JJ) (talk) 23:39, 3 August 2016 (UTC)

IP202's work here is largely paraphrased from Uyeda et al 2008, but the Uyeda paper provides additional references and elaboration. I don't believe either IP202 or Uyeda meant their remarks to be interpreted as you're saying. Always assume good intentions. I have edited IP202's section to hopefully clarify what Uyeda was trying to say. JerryRussell (talk) 04:04, 4 August 2016 (UTC)

"Good intentions" is not a factor here, other than someone being here only to promote something (see WP:NOT#SOAP) is a lack of good intentions.

Before you attempt to "clarify" what Uyeda said, check out Uyeda's actual words in his 2013 IWEP presentation.

Just to whet your appetite, I offer some tidbits. The red and bolding are per the original.

• Most precursors are non-seismic, which means cannot be predicted by seismometers.

This is a fetish of his, that seismologists work only with seismometers.

He also attributes the failure to make any predictions in Japan to:

• seismologists kept monopolizing the program, knowing seismology was unfit.

A key theme:

• "EQ Prediction Village" A bureaucrat-academia-industry Complex

Lots of scary disaster pictures, and exclamation marks, then:

• Thus, the "EQ Prediction Village" was established, with guaranteed funds free from burden of making any prediction.
• The further decided that precursor finding is impossible for anybody.
• Therefore, precursor research is NOT science.

But wait, there's more!

• There are signs that EQ prediction is possible by ordinary "science", only if non-seismic precursor search is fully employed.

• In particular, electromagnetic & geochemical phenomena are promising because many successful predictions have actually been achieved.

Gee, who did that?

All of this is very dubious. Do check the source.

See also similar comments in his 2007 Personal View on Short-term EQ Prediction.

I reiterate that the POV I have identified in the "Classification" edits is the very fringe views of Uyeda, a foremost proponent of VAN. And that those edits are inaccurate. ~ J. Johnson (JJ) (talk) 23:28, 4 August 2016 (UTC)

Hi JJ, thanks for the link to the Uyeda presentation. He does present a different slant from when he is writing for peer review. One huge difference is that Uyeda et al 2008 does discuss long-term prediction as a type of prediction, while the slide show denies that this is any type of prediction. I tend to agree with Uyeda's slide presentation, that it's more useful to speak of long-term forecasts instead of predictions. But both terms are valid English syntax, and both are in use, so perhaps our article should avoid taking a dogmatic position.

Uyeda's disdain for seismologists comes through loud and clear in the presentation, while it is very muted, almost subliminal, in the peer reviewed journal paper.

Isn't it true, though, that most seismological monitoring stations don't include EM sensors of any kind? This is the point Uyeda is making in the paper: he's complaining primarily about seismic instrumentation, and only incidentally about seismologists.

Uyeda's concluding remark that earthquake prediction is moving to the private sector, is fascinating. I would suspect that this must be true. I don't know whether to think this is really a promising growth industry, or simply a way to separate credulous investors from their money.

You say: "...those edits are inaccurate." Do you mean IP202's edits? Did my edits help the situation, or is the section still badly POV in your view? I'm not necessarily arguing to keep it, it's almost redundant to the information in the lede paragraph. But I hope I've made it accurate and neutral. JerryRussell (talk) 00:46, 5 August 2016 (UTC)

Uyeda's point is not that "seismological monitoring stations" (meaning any hole in the ground with a seismometer?) "don't include EM sensors of any kind", it's that seismologists are in charge. But it's all .. well, I was going to call it bullshit, only it's more like the industrial waste that was being sold to farmers a few years back as fertilizer. Uyeda's argument is basically that seismology grabbed bunches and bunches of money but yielded no results, so isn't it time to give money to "non-seismic" research? (Probably meaning the VAN stations he was running till about 2012.) Only it's not like he didn't get any money: in 1995 Japan invested 250 yen (about US$3 million at the time) to set up 20 VAN type detecting stations. In addition they gave Yeda about 30 million yen to "revitalize" VAN research. Mind, this was in 1995.

When Uyeda goes on about seismologists not finding any short-term precursors, and implying that VAN could do better, he is being dishonest, because (in Japan at least) VAN could NOT do better: they also made zero predictions. ~ J. Johnson (JJ) (talk) 23:36, 6 August 2016 (UTC)

Uyeda et al 2008 did mention an excuse for the failure of the VAN program in Japan: he says they set up the monitoring stations in the area of the much-feared Tokai earthquake, which never yet happened in spite of being forecast for the last 20 years; while the big earthquakes that have actually occurred, have all been in areas where no VAN monitoring stations were built. Cry me a river, I know, but that's his story. JerryRussell (talk) 00:14, 7 August 2016 (UTC)

Fact: In 1997 alone, the USGS lists M5.6, 5.8 and 5.9 events in the Tokai region. If you include the Tonankai area (100-200 km away, well within the margins in other VAN analyses, there are a number of events M>5 (and even M>6 and M>7) throughout the late 1990's and early 2000's. Shouldn't these be significant enough? Elriana (talk) 21:40, 12 August 2016 (UTC)

The "Personal Views" essay is more realistic and introspective. "One can blame the biased views of outsiders, but part of this situation is certainly due to inadequateness on our side, namely our science has not been convincing enough." I find this humility much more charming than the bluster in the slide presentation, or the formality of Uyeda's peer reviewed presentations. JerryRussell (talk) 00:58, 5 August 2016 (UTC)

It seems we 've been missing my post to Jerry's sandbox talk page since August 1: Political decisions impact on short term earthquake prediction - Academy of Japan - 2013 ("the government will never make short-term predictions") --IP202-46.198.213.62 (talk) 07:30, 5 August 2016 (UTC)

Yes, I'm sorry I missed that. Uyeda says in that paper also, that government cannot make short-term predictions because the consequences of any error are far too grave. Yet there is tremendous demand for predictions, so he expects that the private sector to step into the gap. JerryRussell (talk) 18:40, 5 August 2016 (UTC)

It's not that governments "cannot make short-term predictions...", it's a problem with making any kind of potentially catastrophic prediction, and it's a problem for any entity. Remember that failure to predict (whether not known, or known/suspected but not announced) is also problem. Or to predict (give the alarm) when nothing happens. (See chart in the Earthquake prediction#Evaluating earthquake predictions. See also the note [currently #21] about China introducing "tough regulations intended to stamp out 'false' earthquake warnings....") As I have just commented above, VAN is probably happy to give their predictions to the government, and let the government take the heat for publishing, or not publishing.

An earlier version of this article had more on this, including a link to the CEPEC's notification protocol. ~ J. Johnson (JJ) (talk) 23:11, 6 August 2016 (UTC)

But at the same time (if the sources are all correct) the Greek gov't set up an institute for the specific purpose of receiving VAN predictions, and Papadopoulos is complaining that VAN is not submitting predictions to them. Is VAN still getting gov't grants in spite of submitting their reports to ArXiv and to the press as they deem appropriate, rather than to the gov't? Or have they gone commercial / rogue? Curiouser and curiouser. JerryRussell (talk) 23:31, 6 August 2016 (UTC)

CEPEC notification protocol? Hmm... OK, here's one source. This is already in the source list for the article, though not referenced anywhere. Jordan & Jones 2010 [1] talks about CEPEC having issued a short-term forecast of a 1 to 5 percent chance of a major San Andreas Fault earthquake in late March 2009. JerryRussell (talk) 00:44, 7 August 2016 (UTC)

Jerry, I am a bit put out by your last edit, with the summary "Folded "classification" section into lede. Separate section was undue weight, and POV, see talk." First, that the "Classfication" section was separate was not an issue, and not why it was POV and UNDUE, so folding it into the lead was not a cure. And by mixing the material it made it harder to identify and tag or remove. Second, even if it was not POV and UNDUE, folding it into the lede is NOT a good idea. The lede (lead) is supposed to be a bare summary or introduction to the whole article, and is not the place to hash out details such as specific claims of "statistically significant faults", or even "Geller et al. are skeptical ...."

Third, and very importantly: where you changed "Earthquake prediction is sometimes distinguished from earthquake forecasting ..." to "Short-term earthquake prediction is sometimes distinguished from long-term or intermediate-term prediction, which is also known as earthquake forecasting", you altered material attributed to Kanamori (2003). And where you wrote that "Long-term prediction can be defined as the probabilistic assessment of general earthquake hazard ...", well, that is just wrong. These are the particular grounds why I will be reverting your edit.

I also object to any space spent on "classification" of short, intermediate, or long term predictions. Quite aside from whether Hayakawa's particular classification is meaningful, these distinctions are unuseful. ~ J. Johnson (JJ) (talk) 22:23, 9 August 2016 (UTC)

P.S. Having reverted Jerry's edit, I partially restored it to the extent of deleting the "Classification" section. If anyone objects just put back in. And this time I'll tag it, and we can continue thrashing it out. ~ J. Johnson (JJ) (talk) 22:33, 9 August 2016 (UTC)

OK. I'm sorry you didn't like my attempt to merge the long-term vs. intermediate-term distinction into the earthquake forecasting and general probabilistic assessment material from Kanamori. I thought what I said would be a non-controversial expression of the concepts from the references, but apparently not. Bygones.

A lot of the material in the Classification section, as it stands, is just repetitive of material in the lede. The distinction between intermediate-term and short-term prediction is sourced, and if intermediate-term prediction actually can work pretty well (as claimed by Kossobokov) then perhaps the classification is useful, and the information is important enoough to belong in the lede. From an insurance perspective, I think a 10-yr vs. 100-yr distinction is pretty clearly useful: 10 years is within the economic life of most buildings, but 100 years is beyond anyone's planning horizon.

We could also consider moving the Kossobokov reference down to the seismicity patterns section, moving the Uyeda quote and discussion to the EM precursor section (or deleting it), and briefly mentioning the intermediate vs. long-term distinction somewhere.

Discussion continues! :-) JerryRussell (talk) 23:39, 9 August 2016 (UTC)

The entire "Classification" section (added by IP*202) is suspect. It is attributed to Uyeda, Nagao & Kamogawa (2009), and Hayakawa (2015; "preface & pages 9,10"). But the classification is not Hayakawa's, it is entirely from Uyeda et al. And we should regard anything by Uyeda as dubious. E.g., while the text only says that "other practitioners such as Kossobokov[11] claim to have achieved statistically significant results...", what Uyeda et al. really said is: "intermediate-term prediction of large EQs world-wide is already in the statistically proven stage (e g., Kossobokov et al., 1999)." Basically, a single claim from the 1990s. To even mention this single claim, when later and more authoritative sources deny any such claims, is a violation of WEIGHT.

Another example: to characterize the scale of "short term prediction" takes only one short sentence. The rest of that section is just a coathook for promoting Uyeda's argument as to why short-term prediction has failed.

I could dispute other aspects of that section (such as the "methods used"), but (having other work to do) I will mention only that this section is not useful. Yes, predictions are sometimes referred to as short or long term, but these are relative terms, and categorization on such a basis is entirely arbitrary. This section adds little if anything to the article, as these terms are not really useful, and any necessity for explaining such terms can be done in a single sentence. ~ J. Johnson (JJ) (talk) 22:30, 11 August 2016 (UTC)

Hi JJ, I agree the section has some problems. Since you're busy, I'll make another attempt to fix it. But, I'm curious as to who are these "more authoritive" sources who have attempted to refute the claims in Kossobokov et al 2011? Vladimir Keilis-Borok is on the author list. His Wiki bio page CV seems impressive enough, although it also mentions that his ability to make earthquake predictions have been questioned by USGS and CEPEC. The USGS citation appears to be a link to a sort of blog page, and the CEPEC link is dead. That sort of nameless bureaucratic criticism hardly seems on a par with peer reviewed publications. Maybe you know of something else? JerryRussell (talk) 22:54, 11 August 2016 (UTC)

In our Wiki article, the section on M8 and Keilis-Borok says RTP is the same thing as M8, and it gives as critical citations: Zechar's PhD dissertation, and a popular book. Seriously, this is supposed to be definitive proof that Keilis-Borok is bogus? This paper [2] seems to be a better reference for Zechar, but I'm not even sure RTP is the same thing as M8/MSc. JerryRussell (talk) 23:29, 11 August 2016 (UTC) tweaked JerryRussell (talk) 04:24, 12 August 2016 (UTC)

Who said anything about Kellis-Borok being "bogus"? That's purely an over-reaction on your part. Also, I wonder if you aren't taking quite enough time to consider what you are reading in the text. The text does not say (as you state) that "RTP is the same thing as M8"; it quite clearly (no?) describes the M8 family of algorithms as including the RTP method. It did not (previously) mention MSc because that is not so much a separate method as an extension of M8. If you have any questions on that please consult the cited source, Tiampo & Shcherbakov 2012, §2.4.

Tiampo & Shcherbakov also explain how the statistical reliability claimed for M8 is attenuated so that the probability gain is "generally small". The ICEF report also discusses (p. 337) Kossobokov's testing of the M8, MSc, M8S, and CN algorithms, and also explains how "large alarm areas, high error rates (e.g., 30-70% false alarms), and relatively low probability gains limit the practical utility of these methods as deterministic prediction tools." P.S. For more detailed criticism see Kagan & Jackson, 2006.

I don't know what "USGS citation" you are referring to, but the link to the CEPEC paper is still good (I just downloaded it). Your "nameless bureaucratic criticism" is uncalled for; the CEPEC (California Earthquake Prediction Evaluation Council) is the "bureau" of the California state government charged with evaluating earthquake predictions.

BTW, the "popular book" you refer to is by a noted expert. It is a secondary source of the highest quality, and much more palatable than most peer-reviewed sources. ~ J. Johnson (JJ) (talk) 22:36, 12 August 2016 (UTC)

Hi JJ, I'm glad you're not saying that Keilis-Borok is 'bogus'. Regarding Uyeda's claim that Kossobokov & K-B methods had achieved statistical significance, you did complain that later and more authoritative sources deny any such claims. It seems very possible to me that M8's probability gain might be relatively small by some criteria, and yet also statistically significant. And I still question whether USGS, CEPEC, Zechar or Hough should be described as "more authoritative" than Keilis-Borok.

I think there's some confusion about RTP. I looked at Tiampo & Shcherbakov 2012, §2.4 which mentions M8, MSc, ROC and Accord methods from Keilis-Borok, and an RTL (Region-Time-Length) method by Sobolev & Tyupkin, who seem to be part of the same research group. RTP (Reverse Tracing of Precursors) isn't mentioned in Tiampo & Scherbakov, but is described by Keilis-Borok here:

https://www.math.purdue.edu/~agabriel/rtp.pdf

In this paper, RTP is described as a "new approach" rather than part of the M8 family. It's the RTP method rather than M8 that is analyzed in Zechar. Maybe RTP doesn't work even as well as M8. JerryRussell (talk) 23:34, 12 August 2016 (UTC)

I'd be happier if you would strike your entire "Seriously, this is supposed to be definitive proof that Keilis-Borok is bogus?" as a red herring. Likewise for "the CEPEC link is dead" (untrue), and "nameless bureaucratic criticism" (uncalled for). As to "USGS", that covers a LOT of gound. If you are going to wave that around please be so good as to specify which citation you are talking about.

Kellis-Borok's career focus was trying to predict earthquakes by statistical algorithms of various kinds. While there was some research where they claimed that the numbers showed some statistical significance, the overall judgment was that these predictions were too vague to be of any use. On this I consider the ICEF to be authoritative. Zechar (in several papers, including his dissertation) and also Kagan & Jackson are also very persuasive, as they are very specific in their criticism. ~ J. Johnson (JJ) (talk) 22:35, 13 August 2016 (UTC)

OK JJ, I struck the "bogus" comment. The CEPEC link I was referring to is on the Wiki bio page for Keilis-Borok, and it is still dead as of now. The USGS link I was referring to, likewise, appears on that bio page. It points to an unsigned essay that reads a lot like it was written as a summary of a committee discussion. Accordingly I feel comfortable with my "nameless" criticism.

As to whether Keilis-Borok predictions are "useful", wouldn't that depend on what the intended use is?

Here is the text from the Wiki bio page. The links in question are given as embedded links rather than in citation form, so this really does need fixed:

His team of researchers have used new algorithmic methods for earthquake prediction. Keilis-Borok's method has been retroactively applied to 31 cases dating back to 1989, with correlation 25 times (not including two near misses), including the Samoa area quake (September, 2009) and the Sumatra quake (September, 2009). In response to his prediction of an earthquake in California in 2005, US Geological Survey has said: "The work of the Keilis-Borok team is a legitimate approach to earthquake prediction research. However, the method is unproven, and it will take much additional study, and many additional trial predictions, before it can be shown whether it works, and how well." [3]. The California Earthquake Prediction Evaluation Council determined, "To date there is no evidence that these, or related methods, yield useful intermediate term forecasts." [4] No earthquake occurred in the predicted location or time period.

JerryRussell (talk) 00:24, 14 August 2016 (UTC)

Thank you for striking the abandoned comment, and for clarifying your other references. And let's note the dead links you refer to are from the Vladimir Keilis-Borok (which is seriously deficient in several respects), not anything linked to in this article or discussion, so I don't see how they are relevant here. I will point out that while (whether you like it or not) the USGS blurb on the K-B prediction is accurate, it is not used anywhere in this article or discussion.

Now let me ask: what are we discussing here? I opened this discussion stating that the "Classification" section (now gone) and the material in it (some which remains in "Prediction methods") was non-NPOV (and pointless), pointing particularly to the claim of "statistically proven". That claim was attributed to Kossobokov et al., who were testing the M8 algorithm of Kellis-Borok. In the course of this discussion I have shown how the statistical significance of this claim is of little practical utility because the large alarm areas, etc. "As to whether Keilis-Borok predictions are "useful" ..." — has anyone said they are useful?

Regarding the "Classification" material now in "Prediction methods" I will point out right off: Intermediate-term and long-term predictions are not "also known as forecasts". That is false, and contrary to the sourced explanation provided in the lede. Also incorrect is the statement that research on short-term prediction "is focused on finding precursors." And I question whether this short/intermediate/long term distinction really warrants any mention. ~ J. Johnson (JJ) (talk) 23:58, 14 August 2016 (UTC)

Hi JJ, you ask: "what are we discussing here?" What I'm trying to do is to develop text for the article that will satisfy both you and IP202 (as well as being, generally, accurate). I can only assume that IP202 thinks this material is important, since he inserted it. Yet you are apparently also still dissatisfied, since the conversation continues. IP202 gave two sources who specifically mention the time scale distinction, and Keilis-Borok explicitly describes his work as "intermediate term", so it does seem to be a term that is employed in the art, and is found useful by some practitioners. Accordingly, I think it's worth some space in the article. I agree that giving an entire section to the subject was undue and served as coat-rack.

Per your recommendation, I've deleted the claim that long-term and intermediate-term prediction is the same as forecasting, although I'm not sure what the difference is. Also, do you know anyone who is making short-term predictions as defined in the article, with time scales of a few weeks, without using precursors? Uyeda made this peer reviewed statement that precursors are necessary for short term prediction, and I don't see anyone contradicting that.

Since this is an article about earthquake prediction, my assumption is that anyone reading the article finds the topic scientifically interesting, whether the results are useful or not. It also seems clear enough to me that any earthquake prediction method that achieves statistical significance, is going to be useful for some applications, such as setting insurance policy rates. The complaints about "usefulness" of M8, I interpret in the context of someone looking for an evacuation warning, which M8 obviously is not. I have no problem if the article mentions the sourced complaints that M8 is not "useful", but do you agree that the article should also mention that it is is statistically significant? This is the explicit claim made by Uyeda and Kossobokov, and I haven't seen any claims to the contrary. JerryRussell (talk) 02:18, 15 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Suggestions re citations

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Jerry, and anyone else adding citations:

1. {{Citation}} does not need an explicit |ref=CITEREF... (unless there is some special case thats needs handling), nor even |ref=harv; suitable CITEREFs are generated automagically.

2. The {{citation}} template – unlike the {{cite xxx}} templates – does not automatically add a period at the end; this needs to be added manually. (A tip: I usually leave the terminal period off until I have inspected the result and tested any links.)

3. I strongly recommend that the closing pair of braces be put at the beginning of a line. I have considerable experience in this kind of formatting, and tacking the double-brace on the end of a long line (like the idiot doi-bot does) makes it harder to parse one citation from the next, or even to see that there is a closing double-brace.

4. I also recommend (somewhat strongly) putting a space before all vertical bars ("|"), and a space after all equal signs, as this best distinguishes different parameters, and makes the data standout from the labels.

5. Dates should be in "DMY" format. E.g.: "2 August 2016".

Thank you. ~ J. Johnson (JJ) (talk) 23:51, 3 August 2016 (UTC)

OK, fixed for the one citation I did so far. If you don't mind, I'll wait to do the citations for the disputed paragraph about 2008 event, until we decide if that paragraph survives. JerryRussell (talk) 01:30, 4 August 2016 (UTC)

Good. And I just fixed some others. (Though I did not try to make them perfect.) And sure, no rush to perfect what isn't going to stay. ~ J. Johnson (JJ) (talk) 21:43, 4 August 2016 (UTC)

I tried to turn Jerry's citations to harv style. {{Harvnb|Uyeda|Kamogawa|2008}} citation is missing. --IP202-46.198.213.62 (talk) 21:33, 6 August 2016 (UTC)

Thanks IP202, I got the missing citation. Looking good now, I think. JerryRussell (talk) 23:01, 6 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Libel on VAN 1983-1995

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

• The section is using biased wording. "Closer examination shows a different story" gives an aroma that sure not is encyclopedic tone, and this aroma is left behind overall the section. "it is impossible to verify that unsuccessful telegrams were not quietly discarded" is being stated with a confidence that VAN group are liars. These are just examples, half the sentenses or paragraphs follow this line.

• The selection of the two examples has been done with unjustified criteria, using only secondary sourced material. These examples are equal to examples of success that are omitted. Third sources presenting the debates and the overall review of VAN method are really important, not fragmented material. The size of the section should be less than half and there is still not cut-down.

• The scientist who does not act like one, stating that resources should not be given to science to save people because those killed by earthquakes are not Greeks and Greeks have other priorities on casualties, does not really belong in the VAN section, even if he refers to it. There is going to be a different section that has to do with political acts, puppets, wording and decisions, beginning with Uyeda review criticizing the Japan government, published in the peer-reviewed journal of Japan Academy. The comment on "killing the dead frog" with this addition, is that this saying indeed has to stay inside the article, only under a different section, as inside VAN it only adds to the libel.

• There are still two missing published review statements, which were removed without justification, "The public impact of VAN's predictions has been large because lives have actually been saved at some disastrous earthquakes" / "VAN has well survived the test of time" (Uyeda, Nagao, Kamogawa 2009) and that VAN method is still "very controversial" (Papadopoulos 2011).

--IP202-94.66.56.26 (talk) 01:04, 15 August 2016 (UTC)

Hi IP202, I agree "Closer examination..." appears to draw a conclusion about the controversy in Wiki voice. As to the claim "it is impossible to verify...": That is a sourced quotation from Geller, and furthermore it seems to be factually accurate. If VAN was sending telegrams only to close associates, there would be no way for outsiders to verify. Considering the stakes, it seems to be a valid concern to be expressed. At the same time, there is no evidence that any telegrams were in fact discarded, aside from the lack of consecutive numbering. Non-consecutive numbering could have innocent explanation. So I personally would support removing the insinuation even though it's sourced. I understand there's ArbCom guidance that says not to delete long-standing sourced material without obtaining consensus on talk page first.

The selection of two examples "such as (typically) a prediction of an earthquake at either 300 km "N.W" of Athens, or 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit" is intended, as it says, to be a typical representation of problems with VAN predictions, namely a lack of precision. The selection could be criticized as "Original Research" but then again it might be equivalent to an obvious application of simple arithmetic. I would have no objection if you were to pull out a couple of examples of 'successful' predictions in which the location, time and amplitude of the quake were all predicted accurately in advance. On the other hand, I would also have no objection if someone were to strike the text about the two predictions as WP:OR, but it's better if we can agree with JJ here in the talk section so that we don't get into edit warring.

The article already has a section 'evaluating earthquake predictions' which seems like a good place to talk about items such as the Stiros article, and perhaps the Uyeda review calling for more funding of precursor research. I agree that Stiros did not single out VAN as the only waste of money in EQ research. I took the BOLD tactic of moving the Stiros material there and elaborating on it, we'll see if it sticks. JJ will oppose including the Uyeda remarks, but I would support it.

The idea that VAN has saved lives is already discussed in the article. Statements that VAN 'has survived the test of time', or conversely that it is still 'controversial', seem basically very subjective fluff to me, and not particularly encyclopedic.

Agree that the VAN section is too long, but I think the first step is to get it to the point where everything in it is neutral and accurate. Then we can look at cutting material, or re-phrasing for length reduction.

Are you the same editor as 94.66.56.77? I do not support the statement that VAN submissions to ArXiv are always simultaneously submitted to peer reviewed journals. From what I can see, many of the ArXiv submissions are multiple editions of the same article, and I doubt that the many revisions would be reflected through into journal editions. I do not support the statement that a "prediction" "warning" was issued through Ethnos, since the crucial information about the location was admittedly and intentionally omitted from that publication. JerryRussell (talk) 03:17, 15 August 2016 (UTC)

What, "defamation" again? Did you think we had already forgotten the discussion above (hardly two weeks past) at #The neutrality tag: update and prognosis? Please remember that the essential element of libel (and slander, the other form of defamation) is a false statement that damages someone's reputation. Glancing over your comments, it seems that my comment of 31 July is still applicable: "there has been no showing here of any false or slanderous statement".

You object to "Closer examination shows a different story" because it – gives an aroma? This statement transitions from the grand and totally unsupported assertions by Varotsos and Uyeda that VAN predictions have saved lives to the closer examination showing a different story, that there is no grounds for making such a statement. You also object that the statement about the possibility of unsuccessful telegrams being "quietly discarded" is "being stated with a confidence that VAN group are liars." I am not aware that "stated with confidence" is any kind of offense (in law, or at WP), and there certainly is no statement that VAN group are liars. (No? Then point to it.) Your objections are entirely unfounded.

In your third section you seem to be saying that having Stiros' quote in the VAN section "only adds to the libel". But, again, you haven't shown how it is false, let alone how it is false in the VAN section but not elsewhere; you have failed to demonstrate any libel. At best you can only say that such views are embarrassing to VAN, which goes to show your continued biased editing in always burnishing VAN, and muting any criticism.

I remind you of Elriana's comment of 30 July that "[a]ccusing anyone of slander is a serious matter and contributes nothing but a dismissive and contentious tone to this discussion." Your edits and discussion are tendentious.

Comment: the VAN section has gotten longer because a certain editor keeps challenging what was originally summarized, requiring fuller explanations and citations. ~ J. Johnson (JJ) (talk) 04:17, 15 August 2016 (UTC)

Hi JJ, in IP202's defense, it is interesting to realize that defamation law in Greece is very different from the USA. In some cases, it is illegal to make insulting statements regardless of whether they are true or not! see [5] As of 2010, however, US defendants are protected by the SPEECH Act from defamation actions in foreign countries in most cases. So, IP202, please understand that Wiki is not bound to follow Greek libel standards.

IP202: by that same token, according to US defamation law, Verotsos and the VAN would most likely be considered public figures by virtue of their notability in the press. Accordingly, to reach the level of actionable libel, a plaintiff would need to demonstrate actual malice (reckless and intentional disregard for the truth) in order to prevail. And furthermore, there's no evidence of actual malice on JJ's part or anyone else, and your continuing to draw attention to the concern is equivalent to an insinuation of misconduct. Have you read WP:BOOMERANG? If there's ever a DR process involving conduct issues, your own conduct will come under scrutiny.

So I'd like to ask everyone to focus on complying with Wiki policies for this article, rather than continuing to raise defamation concerns. IP202, I think it would be wise on your part to strike out and revise the title of this section.

To frame the issues in terms of Wiki policies -- IP202 believes that the formulation "Closer examination shows a different story" is biased and is not encyclopedic; in other words, a violation of NPOV. Similarly, he believes it is biased to repeat ancient allegations of fraud that have never been demonstrated true, even if they have also never been conclusively refuted. I tend to agree, there is plenty in the printed record of VAN activities for evaluation, without needing this unproven Geller insinuation.

Also, the selection of two examples from a huge list seems to be a possible violation of WP:OR. I suggested that maybe IP202 could select a couple of counter-examples, but that might be just more OR, and also lengthen an article section that's already too long. How do you think this should be resolved? JerryRussell (talk) 16:03, 15 August 2016 (UTC)

Jerry, it doesn't matter what jurisdiction you want to consider, the essential element of defamation is a false statement that damages another's reputation. And THERE HAS BEEN NO SHOWING HERE OF ANY FALSE STATEMENT. That fact alone should be sufficient to quash this complaint of "libel". If such a baseless accusation is repeated (again!) I think a range block would be in order.

IP202 complains of "an aroma that sure is not encylcopedic in style". I suggest that the "aroma" is in the nature of VAN and their work. It has been a frequent criticism that VAN fall short of scientific expectations. That, upon closer examination, VAN's reputation is shown to be damaged, or that their claims of saving lives are unfounded, is truthful report, and we are not required to conceal or perfume it. Indeed, to not show the other side of those claims would be bias.

Jerry, I must take issue with your "unproven Geller insinuation" language. This is one of the "scientific expectations" I just referred to. Specifically, this is Allen's sixth requirement of predictions, that they be published in a form that gives failures the same visibility. (See note 1.) This is an extremely important point (it was discussed in a section since removed), and it is in fact the hallmark of prediction charlatans (in any field) to publish only their successes. It is not Geller's duty to prove that VAN's lists of predictions are incomplete (which, as he said, is "impossible to verify", but it is inferred by the numbering gaps), it is a requirement laid on those who would claim success in prediction to show their record is complete. As VAN have never done so, they have failed the scientific expectations for prediction. As to "repeating ancient allegations of fraud that have never been demonstrated true," I have no idea what you are talking about. ~ J. Johnson (JJ) (talk) 23:26, 15 August 2016 (UTC)

P.S. Jerry, regarding this edit: contrary to your edit summary, Stiros did single out VAN. ~ J. Johnson (JJ) (talk) 23:41, 15 August 2016 (UTC)

Are the VAN telegrams a complete record of attempted predictions?

New subsection, reflecting a slide into the question of whether the VAN telegrams are a complete record, or might be skewed towards publishing/claiming successes. ~ J. Johnson (JJ) (talk) 00:03, 18 August 2016 (UTC)

JJ, I looked again at Stiros, and I do believe I misinterpreted him to some extent. He did not express any negative judgment about cost effectiveness of earthquake research or preparedness, aside from prediction research. I revised my paragraph about Stiros, hopefully you'll find it acceptable. Also I agree he did have specific criticisms of VAN, and I would not be opposed to mentioning some of that in the VAN paragraph. But I don't believe he would find any other prediction research program a more acceptable use of research funds.

I've also made an attempt to revise the paragraphs refuting VAN claims that their predictions have been life-saving, putting them in source voice instead of Wiki voice.

If VAN were such charlatans that they would simply deny failed predictions (as inferred by numbering gaps) then shouldn't it have been child's play for them to demonstrate statistical significance of the predictions that were published after the fact? And yet, aren't the critics claiming that there is no significance even for the predictions that VAN has presented? Isn't it possible that numbering gaps were a result of some process such as, for example, an internal note of a signal that was not prominent enough to merit a telegram?

If VAN peer reviewed publications claim that they have in fact published all their predictions including failures as well as successes, and yet Geller points out that VAN's lists of predictions are not consecutively numbered, is he not insinuating that VAN is lying in their peer reviewed publications? And if he is further arguing that, in spite of such jiggering of the results, they still cannot demonstrate statistical significance -- isn't there some sort of boomerang effect, where Geller himself loses credibility in his disparagement of VAN? JerryRussell (talk) 18:25, 16 August 2016 (UTC)

Well, do VAN peer reviewed publications claim that they've published all of their predictions including failures as well as successes, i.e. every telegram? And not publishing failures isn't fraud, it's the file drawer effect. --tronvillain (talk) 21:37, 16 August 2016 (UTC)

Ahaha, maybe we should have a new section for use of the file drawer as an effective prediction method. For that matter, why should Wiki have an article about Penn State child sex abuse scandal when we could just as well have an article about "Clowning Around at Penn State"? [6]

As to whether VAN claims to have published all their prediction telegrams, good question! Maybe IP202 could find us a citation with a clear statement? JerryRussell (talk) 22:40, 16 August 2016 (UTC)

What? The file drawer effect is a well established phenomenon, but I have virtually never seen anyone propose that it constitutes fraud, especially since it may not even be intentional. Perhaps if someone was actually claiming to be publishing every study undertaken (or in this case, prediction made) it might be. And not reporting a crime is completely different, so I don't know why you'd bring up Penn State. --tronvillain (talk) 23:03, 16 August 2016 (UTC)

It's one thing when someone does some work, then decides to not publish because the results are "not interesting enough". But when someone does a bunch of work, or has a bunch of results, and then selectively publishes only the favorable results - I think most scientists would consider that deceitful at the least, and even rising to scientific fraud. (I leave out what the lawyers may think.) But note: no where in the article is there any suggestion of fraud. The point of view taken in the article corresponds to the rather charitable view of the scientific community (at least as far as anyone says in public) that the selectivity that seems evident may be non-deliberate, even unconscious. It has been suggested that VAN's identification of SES - a subjective process - may be enhanced when there have been recent shocks. (A point with statistical support.) However, the point here is not how this might happen, but that VAN has not shown, and indeed, cannot show, that such selectivity has not happened.

As a more specific response to tronvillain: off-hand, it seems to me that three times they have said (in effect) "here are all of our published predictions" for some period. But that misses the point: not only is their notion of "published" somewhat flexible (more on this below), it's still all a posteriori.

Jerry, it's not quite clear what you are saying. (E.g., for whom "should it have been child's play ... to demonstrate statistical significance ..."?) As to your hypothesis of signals "not prominent to warrant a telegram": please note that (at least in the early years) the gaps are in the numbering of the telegrams. If the signal wasn't "prominent enough" to merit a telegram, then why was there a telegram? To be fair, I believe they do claim somewhere that some telegrams were just "advisory". But that only makes it worse, as telegrams are then ambiguous, and the question of whether there even was a prediction is then argued a posteriori.

Jerry, I wonder if I should remind you of Elriana's comment, above. Insinuating that anyone is insinuating, or lying, or is a liar, contributes nothing to this discussion, and amounts to nothing but a red herring. As I just mentioned, I don't believe any critic has been so severe as to accuse VAN of lying. (Though I could quote others who have called them charlatans.) The critics say that VAN did not use a simple method for avoiding selectivity. For you to interpret this as an insinuation of lying, that is, of deliberate falsehood and even moral inturpitude, especially as the critics have suggested other, more benign, sources, is not useful. This is the kind of hyper over-reaction I would expect from the Athenian.

I may have comments re Stiros when I have more time. ~ J. Johnson (JJ) (talk) 23:26, 16 August 2016 (UTC)

Here is the text from the article in question:

To document prediction VAN has relied on telegrams sent mainly to themselves or close colleagues.[134]Approximately forty telegrams have been disclosed, but there is evidence that hundreds of telegrams were issued (and an unknown number of faxes);[135] it is "impossible to verify that unsuccessful telegrams were not quietly discarded."[136]

I can't read that text without seeing an implication that many telegrams were intentionally disregarded because they were failed predictions, which would be an example of "scientific deceit." If the sources argue that this was done on a non-deliberate or unintentional level, perhaps we should make that clear in the article -- although really it doesn't help VAN much, as it implies a level of incompetence that is perhaps even more shocking than 'moral turpitude' would be in a group of PhD scientists.

If we are looking for a truly benign explanation of the extra telegrams, it would be along the lines of (1) clearly marked as non-predictions, or predictions of sub-threshold events; (2) memos about research projects, papers or personal events that were completely unrelated to earthquakes; or (3) I could see an argument that it would be appropriate or interesting to re-evaluate the entire series of telegrams a posteriori based on some objective criteria applied across the entire series (or a subset) in order to optimize the prediction algorithm. This would be especially appropriate if the telegrams were divided into a 'training' set and a 'test' set.

My point about child's play: if I start by generating 300 completely random earthquake predictions, and then pick the 50 of them that come closest to being correct after the fact, wouldn't it automatically follow that my selected 'predictions' would be enormously better than random chance? Especially if the evaluation criteria can also be adjusted after the fact?

I'm feeling a bit amused at the interplay of euphemism, insinuation and misinterpretation in this discussion. I took Tronvillain's comment about "file drawer effect" as an ironic euphemism for deceitful or inept research practices, and felt that the socially correct response was to laugh and point to another example of an amusing euphemism. In the case of Penn State, if you read the linked article, you'll see that numerous eyewitnesses to the child abuse referred to it as "horsing around", and these euphemistic statements were later taken seriously by other investigators. Turns out Tronvillain was really trying to come up with a benign explanation for VAN results, and was puzzled by my reaction.

Now if the "clowning around" analogy becomes clearly understood here, and I were to then say that we are "clowning around" with VAN, then I would be insinuating that we are engaging in inappropriate behavior with them in the shower.

And before I go any further with that image, let me state categorically that I really do believe we're all doing our best to improve the article. I really appreciate the enormous amount of time and effort that have gone into this discussion, and I personally can say I've learned a lot through the process. I'm very sorry if I've given anyone any impression to the contrary.

I do think it's fair to discuss where the article might be making insinuations or accusations, and especially if it's happening unintentionally. It's all too easy to write materials that can be misinterpreted. Case in point, myself. JerryRussell (talk) 02:32, 17 August 2016 (UTC)

PS -- I was curious about the difference between a 'charlatan' and a 'liar', to the extent that I looked it up. Wiki says

A charlatan (also called swindler or mountebank) is a person practising quackery or some similar confidence trick in order to obtain money, fame or other advantages via some form of pretense or deception....In usage, a subtle difference is drawn between the charlatan and other kinds of confidence trickster. The charlatan is usually a salesperson. He does not try to create a personal relationship with his marks, or set up an elaborate hoax using roleplaying. Rather, the person called a charlatan is being accused of resorting to quackery, pseudoscience, or some knowingly employed bogus means of impressing people in order to swindle his victims by selling them worthless nostrums and similar goods or services that will not deliver on the promises made for them. 

Whereas a liar is simply someone who makes false statements intentionally, for any reason. A woman might lie to Nazi stormtroopers at the door, to save her children, to give an example of a lie for a good purpose. To me, that makes 'charlatan' a rather more insulting pejorative than 'liar'. JerryRussell (talk) 02:44, 17 August 2016 (UTC)

Jerry, you are (again!) over-reading the text to see an implication of intentional discarding. Look more carefully: Geller did not say, nor insinuate, that anything was discarded, intentionally or not. He said only that (VAN having failed to address Allen's sixth requirement, as I explained above, and as mentioned in the first footnote of the article) it is impossible to verify that any telegrams were not discarded. (Or that the dog ate them, or whatever.) AS I HAVE SAID BEFORE: it is not Geller's duty to prove that VAN's list of predictions are incomplete, it is a requirement laid on VAN to show their record is complete.

As to "looking for a truly benign explanation of the extra telegrams": why should we be looking for that? Why, when VAN totally screwed up a basic expectation of making predictions, should we have to make excuses for them? But you seem to have missed essential point: the "entire series of telegrams" is no longer knowable, because, just as Geller said, it is no longer possible to verify what is missing. ~ J. Johnson (JJ) (talk) 21:53, 18 August 2016 (UTC)

Reply to: "As to whether VAN claims to have published all their prediction telegrams, good question! Maybe IP202 could find us a citation with a clear statement? JerryRussell (talk) 22:40, 16 August 2016 (UTC)" — Preceding unsigned comment added by 94.66.56.76 (talk)

The abstract of the paper "Basic principles for evaluating an earthquake prediction method, Geophys. Res. Lett. 23, 1295-1298, 1996 by Varotsos et al" which is the first article in the special issue of Geophysical Research Letters (Vol. 23, No.11, May 27, 1996) published under the title “Debate on VAN” (under the editorship of Robert J Geller) starts as follows:

"A three year continuous sample of earthquake predictions based on the observation of Seismic Electric Signals in Greece was published by Varotsos and Lazaridou [1991]. Four independent studies analyzed this sample and concluded that the success rate of the predictions is far beyond chance. On the other hand,..."

More, the captions of Tables 1 and 2 of Varotsos and Lazaridou [1991] (Latest aspects of earthquake Prediction in Greece based on Seismic Electric Signals, Tectonophysics 188, 321-347, 1991) read:

• TABLE 1. Complete list of telegrams issued from May 15, 1988 to August 10, 1989.
• TABLE 2. Complete list of telegrams issued from April 1, 1987 to May15, 1988.

Additionally, Appendix 3 for Table 4 of the paper mentioned above states that it constitutes the "continuation of Table 1", i.e., the five telegrams which were issued between August 10, 1989 and November 30, 1989.

There are clear statements by VAN that the sample of the predictions was continuous as well as that the list of telegrams was complete, but the editor responsible for writing the section claims that it is "impossible to verify that unsuccessful telegrams were not quietly discarded."[136]"]. The citation No [136] is the following paper: Geller, R. J. (1996a), "Short-term earthquake prediction in Greece by seismic electric signals", in Lighthill, J., A Critical Review of VAN, World Scientific, pp. 155–238. In other words the editor of the section (JJ) proceeded to the above unusual claim (which with no doubt is an unacceptable insinuation) solely based on a citation by Geller. --IP202-94.66.56.76 (talk) 18:03, 17 August 2016 (UTC)

P.S.1 Let's avoid Original Research with examples an keep the section readable. P.S.2 The title of the section under discussion only reflects the feeling on reading the article, not any intention for legal actions. --IP202-94.66.56.76 (talk) 18:21, 17 August 2016 (UTC)

Those are the same sources cited in the article. The number of predictions/telegrams is discussed in the footnote that starts by citing Varotsos & Lazaridou 1991, Table 3 (currently note 128). Though Table 1 in VL91 is captioned "Complete list of telegrams issued from May 15, 1988 to August 10, 1989", it is not even that complete, as the last entry is for a telegram issued "23-07-89" (that is, July 23rd). Table 2 is captioned "Complete list of predictions [not telegrams!] issued from April 1, 1987 to May 15, 1988", without any reference to telegrams. In both cases we have only the authors' allegation that these lists are complete; we have no way of knowing whether some predicitons might have been quietly eaten by the dog.

As stated in the article, VAN has tabulated ("published" has multiple meanings here) less than one hundred predictions. Yet, as noted in one of the notes (is it too much to expect the participants here to read the text and notes?), Varotsos et al. 1988 refer to telegrams numbered 364 and 370. So where are the two hundred and some telegrams never seen in public?

Note that the foregoing question is NOT "solely based on a citation by Geller", it is a quite reasonable question likely to occur to any attentive reader of just the VAN sources. And all of these allegations of insinuations are bullshit.

IP202: "Libel" is a legally defined tort, and a potential basis for damages. As you have no basis for claiming libel (nor slander), you certainly should back away. But as you have brought this up several times, without ever showing any support, perhaps you would make a definite, conclusive statement that there is no libel, nor slander, of VAN in the article, and that you will not bother us further with this tendentious claim. ~ J. Johnson (JJ) (talk) 22:01, 18 August 2016 (UTC)

Hi IP202, thanks for the information. Reviewing the citations given for the paragraph, note 135 gives a citation to ICEF and a citation to an article by Papadopoulos. The ICEF section referred to is also by Papadopoulos. The claim in the ICEF report is that between the 1980's and 1999, "hundreds of EPSs [Earthquake Prediction Statements] produced by the VAN team were submitted either to EPPO, to other Greek governmental bodies, or to scientists in Greece and abroad. Many of those EPS's were also announced by Greek or foreign (e.g. French) media..." The other paper makes a similar statement. And, Papadopoulos is contradicting the statement in the Wiki paragraph just above, that the telegrams were sent only to close colleagues. If in fact the telegrams were sent also to various Greek governmental bodies, scientists worldwide, and media outlets, then it seems very possible that of the "hundreds" sent, many were duplicates of the same telegram sent to various locations.

Note 136 is a direct quotation of Geller, it is "impossible to verify that unsuccessful telegrams were not quietly discarded." But, I would think VAN would be taking a huge risk of such a subterfuge being detected, if indeed multiple copies of the telegrams had been sent to multiple recipients including in many cases the press. JerryRussell (talk) 18:56, 17 August 2016 (UTC)

I edited the section to put in source voice and clarify the claims and counter-claims, but IMO it's unlikely that VAN discarded any predictions as charged by Geller. JerryRussell (talk) 19:17, 17 August 2016 (UTC)

How many VAN predictions were posted for the period from 1983 through 1995? Our note 129 says 97 from the VAN article in Lighthill, but elsewhere we had been saying 40. I'm sorry I haven't solved my access problem, I still can't get at all the materials. IP202 or JJ, could you verify how many predictions appear in the VAN tables 1983-1995? JerryRussell (talk) 20:06, 17 August 2016 (UTC)

Attempting to list the VAN "predictions" has numerous challenges. Even basic terms like "VAN prediction" and "published" (or "posted") are ambiguous. The various somewhat overlapping tables are often constrained within ranges of time or magnitude, and even within a given table there can be a question of whether to count predictions, telegrams, or quakes.

BTW, re your latest edit: I think we should stick with Varotsos et al. 1996 (CRV). If you want to add "Jackson & Kagan 1998 say 94", fine.~ J. Johnson (JJ) (talk) 22:07, 18 August 2016 (UTC)

In "Latest aspects of earthquake Prediction in Greece based on Seismic Electric Signals" we have from 1987-04-01 to 1988-05-15 (9 Predictions), from 1988-05-15 to 1989-08-10 (17 Telegrams), from 1989-08-10 to 1989-11-30 (5 Telegrams). That's forty-one telegrams/predictions, plus one isolated telegram on 1990-04-27 (appendix 5). Now "Official earthquake prediction procedure in Greece" (1988, submitted 1987-01-02)) mentions a "telegram No. 364" from 1985-12-19, but that would only be evidence supporting the existence of hundreds of telegrams sent between 1983 and 1985, not between 1987 and 1989. --tronvillain (talk) 20:08, 17 August 2016 (UTC)

And here it says that Table 1 of Varotsos et al [1996] lists 67 predictions, 27 of which were double predictions, for a total of 94 earthquakes. The rate of prediction seems to have really dropped off after 1989, given that at least 41 of those were presumably between 1987-04-01 and 1989-11-30, but as Jackson and Kagan say "Given that such 'predictions' are considered open for up to 2 months, much of the 8 year period was covered. Thus, 'success' by chance is likely, especially since alarms were preferentially issued during heightened seismic activity." --tronvillain (talk) 21:10, 17 August 2016 (UTC)

In the edited section as it is now (many thanks for the improvements), we only need to correct the total number of VAN predictions issued during the period 1987-1995 (probably since only the predictions after 1987 have been considered into the calculations by various independent groups in order to study whether their statistical significance is far beyond chance). This number, as published by Varotsos et al(1996a) - Table 1- is "67", not "97". In the second paragraph we read: In 1996 they published a "Summary of all Predictions issued from January 1st, 1987 to June 15, 1995",[129] amounting to 97 predictions. as well as in the third paragraph: For the period of Jan. 1, 1987 through June 15, 1995, VAN tabulated 97 predictions,[135]. In both the above excerpts the number "97" should be replaced by "67". This also agrees with the number reported by Kagan and Jackson(1998) and by tronvillain just above.--IP202-94.66.56.211 (talk) 16:19, 18 August 2016 (UTC)

IP202, I just saw your comment after updating the article. Jackson & Kagan say that the 67 predictions in the VAN table include 27 double predictions, so the total is 94 predictions. This would be a secondary source assessment of what VAN said, so I think it's what we should report in the article. JerryRussell (talk) 16:43, 18 August 2016 (UTC)

Forgive me for jumping in late, but when VAN issue an prediction, do they specify some kind of range of expected location, time, and magnitude of the earthquake? Isambard Kingdom (talk) 00:10, 18 August 2016 (UTC)

Don't forget that many telegrams have two and even three predictions. Although to cite a different source: NONE of these are predictions! (Ha ha ha, everyone's fooled, the real number is a big, flat zero! No wins, no losses, neither way! :-) ~ JJ 22:13, 18 August 2016‎

@Tronvillain:, thanks for the reference. I updated the article to use the 94 figure and attribute it to Johnson & Kagan, a reliable secondary source. @Isambard Kingdom:, as discussed in Tronvillain's link, a frequent criticism of VAN was that the predictions have not specified these parameters clearly. They give an example: "Significant electrical activity was recorded at IOA-station on August 31 1988 epicenter at N.W. 300 or W. 240 with magnitudes 5,3 and 5,8." The first part of the prediction is simply a report of the SES signal detection; the second part could be a prediction of one or two earthquakes, or it could be a report of earthquakes subsequently observed. The syntax is hard to untangle in this case, but the 'predictions' recorded at ArXiv since 2006 often follow the latter format: the SES signal is noted (now with its amplitude and location and a parameter calculated using Natural Time procedures) and the reader is expected to deduce a prediction from this information. In subsequent updates to the ArXiv report, any subsequent earthquakes are described and associated with the SES signal. JerryRussell (talk) 16:34, 18 August 2016 (UTC)

IK: That has been an issue. Location - often a pair of locations - has usually been given as a distance and direction (e.g., "N.W.") of Athens; a principal problem in trying to evaluate the predictions is determining where the cut-off should be. They have also argued that the boundaries should be stretched when a quake is close. They have claimed that magnitudes are within (plus or minus) 0.7 "M", but there have been issues with which magnitude scale, which catalog, and whether the preliminary or final bulletin should be used for the actual magnitude. In a couple of cases they have disputed the official value.

Their predictions lack expiration times, so there is scope for waiting for a big-enough, near-enough quake to come along. Remember that Greece is very active seismically, even more so than California, and shocks often cluster. IP202 has claimed that VAN's papers rule because they are the latest, no else has criticised them. I think it is more like, given the kind of mess we see above, no one has patience for them. ~ J. Johnson (JJ) (talk) 21:49, 18 August 2016 (UTC)

Hi JJ, you've interspersed several new comments in the threaded discussion above, I'm going to respond to all here.

Given that you acknowledge that VAN papers clearly claim that *all* of the 'predictions' have been disclosed, it is certainly insulting to VAN's integrity as scientists and professionals to claim that they would let the dog eat their telegrams. For various technical legal reasons, this is not an actionable libel in the USA, but IMO you should consider forgiving IP202 for taking offense. As I explained before, if we were all living in Greece it might very well be actionable.

As to your point "It is not Geller's duty to prove that VAN's list of predictions are incomplete, it is a requirement laid on VAN to show their record is complete." My understanding is that it's considered appropriate for scientists to allow their data to be reviewed by their journal review editors based on reasonable doubts, but it's not considered reasonable for just anybody to get access to lab notebooks on demand. In this case, VAN sent the telegrams to several destinations including some independent ones, so any subterfuge could have been detected, as I said earlier.

It might help to see some broader context for the Geller quote. If he's just making some general theoretical point about scientific skepticism, and the value of replication of results by other research groups, while making it clear that he's not specifically accusing VAN of scientific fraud, then I would agree he had a valid point. But by the time we finish explaining all that in our article, it would be more TLDR than it is already.

With the understanding that few if any of the VAN communications are actually "predictions" according to your preferred definition, I find myself more inclined to cut them some slack. Perhaps their communiques should be interpreted as "hey guys, we saw this signal on our equipment, it's been correlated with earthquakes in the past, please consider taking whatever you feel is appropriate action based on this information." Then after the fact they analyze their data to determine what the correlation is. Over time, and with more data, their interpretation has changed.

With 67 telegrams and 94(?) predictions and only 10 hits by their own count, certainly VAN would have to admit to a very high false alarm rate. Hopefully even IP202 will agree it was a serious issue with their work from the 1980's and 1990's. Such a high false alarm rate might not make the work completely useless, but it certainly limits the utility. The lack of precision in the 'predictions' is also an obvious limitation on the utility, to say the least.

I'm a little confused as to what you want to say about the prediction count in the article. I understand that VAN listed 67 telegrams in Varotsos et al 1996, and Jackson & Kagan counted 94 predictions in the table. Do we need to mention the 67 as well as the 94? Again, I was trying for brevity, but I have no objection in principle to spelling out the complete situation.

You wrote: "I think it is more like, given the kind of mess we see above, no one has patience for them." The other side of that coin is, anyone who does see any value in it, has taken the Kool Aide and become a member of the cult! Or in Wiki terms, they're no longer an "independent reputable source". What's a little unusual is, these cult members keep getting their materials through peer review. And in 2008, apparently for the first time in quite awhile, a major Greek newspaper ran with one of their reports; you know what happened next. JerryRussell (talk) 23:55, 18 August 2016 (UTC)

Jerry, please pay better attention, and take more care in your thinking. Strictly speaking, I acknowledge only that VAN claim to have disclosed "complete" lists of - well, just what they have disclosed is a bit murky. Sometimes it is telegrams, sometimes predictions (these are not the same!), and always these are constrained within such varying ranges of time, geography, and magnitude that "complete" loses all definitude.

And PLEASE NOTE: I have not suggested that "the dog ate" any predictions, just as Geller did not suggest that any were "quietly discarded". The POINT is that some form of selectivity, whether in the form of such scenarios as have been suggested, or in a more subtle form, cannot be rejected because VAN failed to address a very fundamental scientific requirement of prediction. If VAN feel embarrassed by this, well, they should, as it is a fundamental failure. But unlike other professionals, when some problem is pointed out, VAN have never shown any willingness to address the problem, but only ignore and deny the criticism, and hector and denigrate the critics.

That mentioning any of this (however embarrassing VAN find it) is "not an actionable libel in the USA" is NOT because of (as you say) "various technical legal reasons". It is because, for lack of an essential element, there is no libel. And what ever the law may be in Greece is entirely irrelevant here, so repeatedly harping on that is a red herring.

As to IP202 taking offense: why should he? Unless, of course, he is a member of VAN, or (as I suspect) very closely asociated, but he has not claimed that.

Your "access to lab notebooks on demand" is another red herring. Allen's sixth requirement for a valid prediction is:

6. It must be written down and presented in some accessible form so that data on failures are as easily obtained as data on successes.

There is absolutely no requirement for access to lab notebooks on demand. Nor is this part of the editorial or peer-review process for an article; it is an essential element of the data collection process.

Regardless of whom received the VAN telegrams, the fact remains: neither VAN, nor any recipient, has provided any list of telegrams sent. However, please don't get too hung up on the telegrams themselves, as that is only one part of the selectivity issue. AS I HAVE ALREADY SAID (23:26, 16 Aug.): It has been suggested that VAN's subjective identification of SES may be enhanced (perhaps unconsciously) when there have been recent shocks. (If you are interested in how professional scientists handle these things look for some of the CSEP documents.)

I think we might agree that an encyclopedia article is not the place to present a fully detailed justification of every point with which someone not here for the encyclopedia might take issue. (Apparently that is what Talk pages are for.) The mainstream views we present are necessarily summarized. Still, in my opinion the article does need the section (now removed) explaining these vital requirements for valid predictions. Lacking that, Geller's comment gives the reader a pithy insight of how selectivity can be an issue. This touches on a broader issue, but it is specifically pertinent to VAN, who reject the principal means of distinguishing scientists from charlatans.

As to cutting any slack for VAN (or their sycophants): why? The mainstream seismology/geology community doesn't, so any such favoring would be bias. Note also that in VAN's earliest reports their method was presented as a verified, operational method of earthquake prediction. It was not then, and still is not, even after thirty years of tinkering.

That second paragraph in the VAN section, on the number of predictions and successes, is way over-detailed, and quite garbled. I think the original was distinctly superior. That it got out of hand is a direct result of trying to accomodate all these nit-picking objections. (Yes, your efforts have been valiant, but summarization is needed. That's why I push much of the detail into the notes.)

On your last paragraph: oh, yes, I fully agree. Indeed, it was similar comments of yours that brought me around to the view that: VAN is pathological science. ~ J. Johnson (JJ) (talk) 01:41, 20 August 2016 (UTC)

If we are going to veer away from the completeness of the VAN telegrams then we should split off another sub-section. Otherwise it's practically impossible to follow the discussion on any particular point. ~ J. Johnson (JJ) (talk) 23:32, 20 August 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Miscellaneous VAN-related NPOV topics

Hi JJ, thanks very much for your comment about my "valiant" efforts. I appreciate the ongoing discussion, and I know the process seems tedious and sometimes repetitive, but I do think we're making progress and I hope you feel the same.

I would have no objection to bringing back a section on Allen's prescription for EQ predictions. Do you have the material handy? It doesn't seem to be in my cache of old deleted sections.

I think both sections on VAN are too long, including the 2nd paragraph. It seems we all agree about that, but we need to be careful not to lose any consensus we've developed, in the process of shrinking it down.

Are there any changes to the 2008 section that you'd like to suggest? I'd like to keep the section, but I'd also like to get rid of the UNDUE tag.

I admit I sometimes trip up by reading things carelessly, but in this case I think your words do amount to an insinuation that VAN's dog might have eaten a telegram or two. ("Insinuate: to suggest or hint slyly; to instill or infuse subtly or artfully, as into the mind...") [7] And I've as much as admitted that I can only rely on VAN's word, as evidence that their tables are complete. I think it would be very much to VAN's benefit, if they would follow Allen's Sixth Commandment. I don't know what more to say about this. JerryRussell (talk) 04:45, 20 August 2016 (UTC)

Well, be more careful!! (Please?) However, I think you really, really, need to get off this "insinuation" idea. It seems that you (and also IP202) have an inadequate appreciation that without proper records (or what ever) a LOT of non-happenings are impossible to verify. VAN should have taken adequate measures, and could have, but even when one method was explicitly suggested Varotsos himself rejected it. Whether "the dog" ate some telegrams, or not, is not the point, The POINT is that THEY CAN'T DISPROVE that, or any of a myriad of other possible forms of selective bias. However much VAN, and IP202, are discomfited by any supposed "insinuations", it remains a FACT: they failed in their experimental procedure, and that in itself is a cause for scientific opprobrium greater than any supposed "libel" that I can "infuse subtly or artfully". All this complaining about "insinuations" is just a stinking, giant RED HERRING.

Re your other comments: how many times to I have say why the VAN section has gotten too long? And, sorry, no, the 2008 prediction section is still UNDUE. It is (as previously discussed) not notable (outside of Greece, at least), nor has it been accepted as valid prediction (let alone, as VAN/IP202 claim) successful, and appears to be included only as a coatrack for "natural time". It is certainly much less notable than other predictions which were previously removed on the grounds of the article being too big. At this point the only change warranted to that section is complete removal.

BTW, IP202's "this" diff (immediately following) does not "help the 2008 section get rid of the UNDUE tag". But beyond that I no more time today to consider his comments. Though in editing I see the use of the word "spectacular". I don't think so. Just more of the hype we see so much of from VAN. ~ J. Johnson (JJ) (talk)

Hi JJ, if the real point is that VAN can't disprove a myriad of possible forms of selective bias, maybe we could find a quote from Geller that makes that exact point, instead of the existing quote that mainly serves to create an unintended insinuation?

Dare I say that the consensus here seems to be in favor of keeping the section on the 2008 EQ and natural time? Each one is sufficiently notable in its own right, and neither is a coatrack for that other. You are making me curious, though, about those other predictions that weren't notable enough to make it into the article.

Did you look at the Rundle et al article? (PDF available here [8], I probably should have included the link in the citation. For earthquakes, we find the 12-month probability for magnitude m>6 earthquakes in California increases from about 30% after the last event, to 40%–50% prior to the next one. What would it take to impress you? Maybe "spectacular" is a bit over the top, but this works for me as evidence VAN has found an interesting approach. JerryRussell (talk) 00:09, 21 August 2016 (UTC)

No "maybe" about it, "spectacular" is definitely over the top. The 'ONLY reason the "2008" section exists is to validate "natural time". It is a WP:COATHOOK, pure and simple. (And Rundle et al. have nothing to say about the event. They only reason for citing them is because they claim to use something similar to "natural time", which is another coathook.)

Look at the content: Of a total of nine sentences, the first four don't even mention the event, they are entirely about "natural time". Sentence #5 tries (albeit pathetically) to establish that VAN are actually making predictions. Sentences 6 and 7 state that there was a prediction, followed by an event, and, almost incidentally, that this event was the strongest in Greece in some 30 years - but none of that has a source cited. But, hey, no problem, because the next sentence, about VAN's report being described in a newspaper, has four citations.

Remove the sentences about "natural time", and trying to establish that there was a valid prediction, there is only part of an uncited sentence to establish any kind of notability. Which is applicable only in Greece.

The event described in the header gets only half the coverage of "natural time". The tail is out front, wagging the rest of the dog. It's a frigging coathook. That you (or even I) should find any of this "interesting" is quite irrelevant. ~ J. Johnson (JJ) (talk) 23:53, 21 August 2016 (UTC)

JJ, I'm sorry I missed this earlier. I have trouble following densely threaded interpolations, they don't necessarily show on my watchlist. I see your point about 'natural time' being largely irrelevant to the topic of this section, and I've tried moving it up to the 'methods' section. As to the rest, the text scrupulously uses the words 'alarm' or 'report' rather than 'prediction'. I agree these are not 'predictions' according to the precise definition used in the literature of this field. JerryRussell (talk) 17:31, 25 August 2016 (UTC)

This would certainly help the 2008 section get rid of the UNDUE tag. In case this is not clear, VAN's natural time is used in mid term earthquake prediction by completely independent groups of top class scientists, the results are so spectacular that they named their enhanced method after natural time and this was published in a first class journal. --IP202-94.66.56.34 (talk) 10:37, 20 August 2016 (UTC)

Jerry, I carefully went through your objective remarks on August 18 (&20) related to the points made previously by JJ. I then read the response of JJ on August 20 which in my opinion is failing to keep any the scientific level I think is proper. I am fully convinced that JJ intentionally proceeded (and still continues while we discuss) to the defamation of VAN research efforts. The following points still reveral his intention should be given attention, in my opinion:

In the second paragraph of the "Electromagnetic variations" section, we read the following two claims:

First claim: "Objections have been raised that the physics of the VAN method is not possible and the analysis of the propagation properties of SES in the Earth’s crust claimed that it would have been impossible for signals with the amplitude reported by VAN to have been transmitted over the several hundred kilometers distances from the epicenter to the monitoring station.[57]" This claim is clearly and precisely answered in the article by Varotsos et al, "Earthquake prediction evaluation standards applied to the VAN method - Reply", Geophys. Res. Lett. 23, 11, 1367-1370, 1996, which was also contained in the "Debate on VAN" published by GRL that was edited by RJ Geller! In particular, in page 1369 of this article an extensive excerpt (covering almost half a page) is written : "As for the question of whether it is possible for the electric fields to be detectable at distances of hundreds of km far away, there is no inconsistency according to the fundamental physical laws. Let us consider ... The latter picture, which allows the SES observation at long distances, consists of a favorable channelling, from the source to an outcrop close to the surface, and an appropriate location of the VAN station at a nearby high resistivity site, that 'amplifies' the electric field. Such a picture also gives a physical explanation of the selectivity effect."

Second claim: "It was also claimed that VAN’s publications do not account for (i.e. identify and eliminate) possible sources of electromagnetic interference (EMI)." This claim is also clearly and precisely answered in the large section (covering almost one and a half page) under the title "Noise reduction" of the article by Varotsos et al, "Summary of the five principles suggested by P. Varotsos et al. (1996) and the additional questions raised in this debate", Geophys. Res. Lett. 23, 11, 1449-1452, 1996. Please note that this article also is contained in the "Debate on VAN" published by GRL that was edited by RJ Geller! In particular, this Section reads: "Noise obstructing the clarity of SES can be classified into three categories: electrochemical, cultural, and magnetotelluric [Varotsos and Alexopoulos [1984a,,b,c].

• Electrochemical noise: It can be easily recognized when a number of parallel short dipoles, for each measuring direction, are installed at every station (Varotsos and Alexopoulos [1984a,b], Varotsos and Lazaridou[1991]).
• Cultural noise: It arises from artificial sources that may lie... the noise is recognized by using a configuration that includes short and long dipoles.. (Varotsos et al [1993a]).
• Magnetotelluric noise(MT). Three procedures are followed by our group:..."

It is really unfair that JJ, although it is absolutely sure that he was aware of the aforementioned answers of the VAN group, he again proceeded to an insinuation that such answers to his claims did not exist. It would be fair to make a little clearer that answers exist.

Moreover, it is also unfair that a systematic defamation against the VAN research efforts exist I also think we should have a parallel look in the "VAN method" article. A separate section has been presented there on "Criticisms of VAN". The claims found there today are untrue because there do exist published answers to the claims but they are not to be found in the article. Scientific argumenting is muted. For example, these criticisms start with: "Currently, the major criticism of the VAN method is that results have not yet been replicated by scientists outside specific research groups in Greece and Japan. Testing of the method needs to be done by scientists unrelated to these two groups. Independent verification is a standard protocol in science." This is simply untrue, because it is being precisely demonstrated, by citing several articles published by various authors in peer reviewed journals, in Chapter 23 of the book "EARTHQUAKE PREDICTION BY SEISMIC ELECTRIC SIGNALS: The success of the VAN method over thirty years" (SPRINGER 2013, ISBN: 978-3-642-24405-60) by Mary Lazaridou-Varotsos, that VAN method has been replicated with encouraging results, beyond Greece and Japan, in various countries including China, Mexico, Italy etc.

--IP202-94.66.56.34 (talk) 15:00, 20 August 2016 (UTC)

IP202, I would repeat my advice to retract your defamation charges, and especially please retract any hint that JJ is acting in bad faith. You've been warned that he is going to take this behavior of yours and ask for a range block.

Thanks for calling my attention to your post above, about Rundle et al. I agree that reference should help establish notability for 'natural time' concept.

I disagree that the article does not currently reply to the claims that physics of VAN method is impossible, and that EMI is not rejected. The structure of the section is: 1st paragraph, VAN claims that SES may be used to predict earthquakes; 2nd and 3rd paragraphs, objections by critics; 4th paragraph, ongoing debate including references to Lighthill, GRL and Uyeda, where further discussion and debate may be found. We could make it a little clearer, but it's not that bad.

We should probably take up your concerns about VAN method article, at that article's talk page. I'll put it on my watch list. Information about VAN replication efforts in China, Mexico, Italy etc. might be helpful for this page as well, but Mary L-V's opinion is not going to be viewed by Wiki standards as an independent RS regarding those replications. JerryRussell (talk) 16:23, 20 August 2016 (UTC)

I am sorry for being that bold, I did feel the boomerang effect and I apologize, I was hijacked by dogs eating telegrams and charlatans and I should not. The article has been indeed improved in JJ's approval so far, so I have striked out and revised around my strikeouts. I understand Mary V-L's opinion is not to be counted, we can only use it as the index that puts together the replication efforts in third countries. I also think it would be fair to give it a try with a clearing in wording in the second paragraph. --IP202-94.66.56.157 (talk) 17:53, 20 August 2016 (UTC)

Thanks IP202, I appreciate your good faith. I've been making some edits this morning, I hope they'll meet your approval. JerryRussell (talk) 18:01, 20 August 2016 (UTC)

L'Aquila EQ neutrality

I've read that the public was reassured by seismologists before the L'Aquila EQ, that there was no evidence of an imminent earthquake. Those seismologists were subsequently convicted of manslaughter. The article insinuates that Giulani's prediction the evening of 5 April was invalid. I wonder if anyone interviewed the 'relatives, friends and colleagues' who were allegedly warned? ICEF is cited as the source for the claim that Giulani's prediction was 'invalid' but I think that's not exactly what ICEF said.

Before leaping into making changes to the article, I thought I'd begin here at the talk page, and ask if anyone else has perceptions about the L'Aquila event. JerryRussell (talk) 17:58, 20 August 2016 (UTC)

You think "invalid" is "not exactly what ICEF said"? Why don't you look at the source and know what they said? Would

At least two of the predictions (on February 17 and March 30) were false alarms. No evidence examined by the Commission indicates that Mr. Giuliani transmitted to the civil authorities a valid prediction of the mainshock before its occurrence.

be sufficient to support "no valid prediction"? Perhaps you don't like the qualification "transmitted to the civil authorities", but what other criteria would you use to determine when mere utterances (to relatives, friends and colleagues, by the thousands) attain the status of a prediction?

Perhaps you would clarify just what point of view is being expressed that you think is not neutral? As it is, the text (the vaunted "WikiVoice") does not directly state that Giuliani's supposed "prediction" was invalid, it only quotes the source. And I had thought this was a reasonable application of another vaunted principle of just laying out the facts and "letting the reader decide". But if that is too much in the nature of an insinuation (I'm getting rather tired of that word), then by all means let's just state directly, based on the source, that two of Giuliani's alarms were outright false, and none were deemed valid predictions.

As to the Italian manslaughter convictions - is that really relevant to to whether Giuliani's "prediction"? ~ J. Johnson (JJ) (talk) 23:44, 20 August 2016 (UTC)

If one wanted to know what Giuliani said to his friends, why not ask them? Did the ICEF interview them? What about the Italian media, or court proceedings for the seismologists' trial? I am curious to get independent verification of what Giuliani said to his friends, and the article doesn't provide it. In source voice, it would be accurate to say "the ICEF determined that there was no evidence that any valid prediction of the mainshock had been transmitted to the authorities before its occurrence." Well of course not, the authorities just publicly humiliated him.

I do think the Italian manslaughter convictions are completely relevant to this section, inasmuch as I seriously doubt that sequence of events would have happened if the seismologists had not been reacting to Giuliani's findings. JerryRussell (talk) 00:24, 21 August 2016 (UTC)

After reviewing the ICEF report more carefully, I see that they're very specific about the accusation that a "valid" prediction was not transmitted to the "civil authorities". Searching for a definition of the phrase "valid prediction" in the ICEF document, I did not find anything specific, but there is a discussion based on "operational fitness" (p. 329) which seems to be related to the precision and specificity of the prediction. The ICEF also casts aspersions against Giuliani's measurements and lab procedures (on p 335) without actually saying what (if anything) was wrong.

The report makes it very clear that agencies DPC, INGV and CGR were responding to Giuliani's false alarms, when they claimed that there is no valid scientific means of short term EQ prediction, and that there was no elevated danger at the time. These were apparently the statements that led to the manslaughter convictions. I believe such statements can be construed as a type of negative EQ prediction, and as such, they definitely should be mentioned in this context.

I would like to do some more reading about this topic. Meanwhile, I made the correction to the source voiced ICEF statement in the article as I suggested above, and I cleared the POV tag. In retrospect, I'm feeling that placing the tag was unnecessarily confrontational, and I apologize. JerryRussell (talk) 16:04, 21 August 2016 (UTC)

Yes, please do more reading. I recommend the articles cited. Start with Squires and Rayner's article in The Telegraph, which pushed this idea that someone HAD ACTUALLY PREDICITED THE QUAKE!!!, but couldn't save any lives because he had been MUZZLED BY THE GOVERNMENT!! An attitude that many people find attractive. But if you look into the other articles (by McIntyre, Hall, Kerr, and Alexander - see the notes, that's what they're there for), you will see that he made two other "predictions", which were false alarms, for which he was told that any future alarms had to through proper channels.

Your latest edit, that Giuliani had not transmitted a valid prediction, leaves the impression that he at least had "a valid prediction". If you want to get more detailed then you should also mention that he made two false alarms. But that is the start of this process by which the VAN section has gotten so large.

The manslaughter convictions are relevant only to the consequences of a botched prediction, and in that regard are already included in the article. (Look harder.) They say nothing about whether Giulani's supposed prediction was valid or succesful. They don't really say anything about prediction per se, unless that it is damned hard to do successfully. ~ J. Johnson (JJ) (talk) 23:55, 21 August 2016 (UTC)

I've been back this morning to do some of the recommended reading. The section on 'prosecution' at the main Wiki article on L'Aquila EQ is pretty good, and I confess I hadn't seen it. The only reference that addresses the science of Giuliani's prediction is the Kerr article, and that's hidden behind a paywall, but I found a bootleg copy. It reviews general mainstream doubt about radon, and quotes Marzocchi as saying that he took a look at Giuliani's material and finds no merit in it. That's not what I would call a thorough and impartial review. The Guardian article is very clear that Giuliani said he anticipated the mainshock and took precautions with his family. Also, everyone agrees that Giuliani gave two false alarms before his successful forecast.

Perhaps a recurrent disconnect that we have, is our view on the value of forecasts that give indications of increased probability of an EQ. I believe such forecasts or alarms are highly valuable, and free individuals can make their own choices about appropriate responses. And if somebody claims to be making 'predictions' when they're really issuing 'forecasts' or 'warnings' or 'alarms', it's just a difference in vocabulary: they haven't read Allen, or maybe they don't agree with him.

One aspect of the Guardian article is very intriguing to me. It says:

The American Geophysical Union had invited him to present his work to its members in San Francisco.

As it turned out, Giuliani's presentation last December went very well. The Americans may not hold a candle to the Italians in matters of disaster management (compare New Orleans to L'Aquila), but they appreciate a free and independent spirit of scientific enquiry. The evidence Giuliani presented aroused intense interest and debate, and the AGU subsequently invited him to take part, with Chapman University and Nasa, in developing a worldwide seismic early warning system.

Chapman University has a department, Center of Excellence in Earth Systems Modeling & Observation, whose charter includes earthquake forecasting as one of four primary activities. Their website states: In particular, there is a real possibility that within a few years, combining large amounts of data and information, particularly from satellites, scientists will be able to forecast earthquakes. The Chapman team works closely with NASA and FEMA scientists and is part of a select group of international scientists who communicate amongst themselves in order to validate earthquake precursor signals, indicating an impending earthquake. JerryRussell (talk) 19:14, 25 August 2016 (UTC)

I've got to get back to work, so a fuller comment will have to wait. But that bit about "a real possibility that within a few years ... scientists will be able to forecast earthquakes", is malarky. (From their mention of "precursor signals" I presume they are using forecast synonymously with prediction.)

Take a look at an earlier version of this article, which shows the article at a fuller state, including 15 notable predictions. Note also the sequence of quotations, tracing out how scientists went from "Earthquake prediction ... appears to be on the verge of practical reality" (1973), to "may be possible within 10 years" (1976), "at least 10 years, perhaps more" (1978), up to "predicting earthquakes is challenging and may be possible in the future" (2012). The topic of earthquake prediction is ever promising - at a superficial level. It is the history, and perhaps most significant lesson, of seismology that: it's not so simple. ~ J. Johnson (JJ) (talk) 23:34, 25 August 2016 (UTC)

People writing grant applications have got to stay optimistic. JerryRussell (talk) 01:07, 26 August 2016 (UTC)

Indeed. But "a promise made is a debt unpaid" which (besides skewing one's assessment of the results) can come back real hard when the funding agency starts asking "where's the beef?" There is a lot to be said for studies that ask "do we have an operational prediction method?" rather than assert "we have an operational prediction method!" ~ J. Johnson (JJ) (talk) 22:51, 26 August 2016 (UTC)

There are some deeper issues to discuss, but we should start with your extended quote from the Guardian. What is the purpose of it, from the perspective of the article? To put Giuliani and his work in a better light, right? But what is this supposed invitation to a respected organization? Did he actually give a presentation in front of a packed auditorium of professionals on the edge of their seats? I remember trying to look into this, but I don't recall finding any evidence of that. And certainly not "intense interest and debate". It is more likely that he was "invited" in the same general invitation made to everyone, and that he merely presented a poster. The reality is that radon has been a favored prospective precursor for a long time, and prior to L'Aquila I doubt he had any scientific reputation outside of the rather small "radon" community. He certainly does not have any publications. What the Guardian published was probably based entirely on Giuliani's statement, and polished on both ends. As to any subsequent invitation to develop "a worldwide seismic early warning system" - well, that is entirely uncredible. Not only is there no evidence of any such invitation, Chapman University - a private, religious college, not noted for any strength in its scientific staff - is hardly in a position to implement any such world-wide system. There are people associated with NASA that have such pretensions, but that is on the basis of EM signals, not radon. The bottom line: the Guardian article is, to a large extent, a puff piece, and not to be relied on. Of course it is intriguing, that's by design, to get people to read it. It differs from rank sensationalism only in degree, and it's not what we do at Wikipedia.

As to this deeper "disconnect that we have", there are two important points to note. First, the prediction/forecast distinction is not "just a difference in vocabulary". Sure, they merge, but there is a BIG difference between saying "there is a 50% chance of a damaging quake within 30 years in this region", and saying more definitely "There's going to be an EARTHQUAKE!" with an implied narrow specificity of MAJOR, HERE, and SOON! Especially when post-event focusing is applied.

Second (and please forgive me if I get a bit agitated here), this idea that "free individuals can make their own choices" is meaningless. Note that this notion is often qualified with "informed". But how can any individual be so informed on all things, and so competent at resolving contradictory information, as to make optimal choices? We can't, therefore we must rely on those who are more informed, and more competent. I.e., experts. So when the mayor of Sulmona was told he had as little as six hours to take appropriate measures, how much time can he take, and then the residents take in turn, in assessing Giuliani's warning?

The ultimate reductive result of "free individuals" making "their own choices" is we just give 'em the raw data. That doesn't work (especially as even the experts can't decide!); raw data almost universally fails to inform, and often leads to panic (e.g., see the Brady predictions), defeating the purpose of prediction. ~ J. Johnson (JJ) (talk) 22:58, 26 August 2016 (UTC)

Hi JJ, I agree that the distinction between prediction and forecasting is important, I'm just saying that there may not be universal agreement on the definitions among all practitioners. I don't fault the mayor of Sulmona for warning the residents. How many people were killed, as a result of that false alarm? How much does it cost to pitch a tent and sleep outside one night, instead of staying inside your ancient unreinforced stone house? The cost of false alarms is vastly overrated. This is why residents of L'Aquila were angry (and, IMO, justifiably so) that Giuliani's later prediction was withheld from them, and that the government told them the risk of an earthquake was minimal and subsiding.

Chapman University's resident earthquake prediction specialist seems to be Dimitar Ouzounov. He published a paper with Gioacchino Giuliani in 2015. So it seems Giuliani is now a published peer reviewed author after all, though the paper is about gamma rays, not radon.

The first paper listed on Ouzounov's CV is Pulinets S.A., Ouzounov D.P., Davidenko D.V. Earthquake Prediction is Possible!? Integral technologies of multiparameter monitoring the geoeffective phenomena within the framework of integrated lithosphere-atmosphere-ionosphere coupling model, Moscow, Trovant Publ., 2014, 144 p (in Russian) so I guess he's another one of those that don't agree with Geller's pessimism. Here's his complete CV which runs to 19 pages and 334 items: [9] Giuliani's paper is item 11.

In addition to the peer reviewed journal, there are five other items of collaboration with Giuliani. Here's a report on radon observations related to L'Aquila April 6 EQ: [10] JerryRussell (talk) 02:54, 27 August 2016 (UTC)

Jerry, what are you trying to say? It seems to me that you have a some what indistinct set of impressions, leading you to think there may be "something interesting" to add to the article. And a number of rather vague questions. All of which I would be happy to explore with you, but we really need some clarification, and a narrower focus. Perhaps on your Talk page?

In this thread you have quoted Dollar's article in the Guardian about the AGU inviting Giuliani to "present" his work, etc. Did you note my earlier comment that this is all puffery, and probably based entirely on Giuliani's saying so?

Re Chapman College, I still say it is not noted for any strength in ts scientific staff. Their top guy seems to be a bit of figurehead, and you will notice that the paper you cited ("first paper on [his] CV") has Pullinets as the lead author. That in itself is a warning flag, as he (and another fellow name of Freund) have been pushing some pet theories about EM signals in the atmosphere, and are rather fringe. Not that there are any peer-reviewed papers saying that in so many words. But UO and OSU both have top-ranked seismologists and geologists; drop in sometime and ask around.

And as I said earlier: the very notion of "a worldwide seismic early warning system" is quite incredible, in that no one has made any system (based on EM, radon, frogs, or whatever) that has predicted earthquakes in any useful sense of "prediction". It is quite unlikely there will be any such system based on radon. ~ J. Johnson (JJ) (talk) 23:36, 1 September 2016 (UTC)

Hi JJ, if you'd prefer to continue the discussion on my talk page, of course I'd be happy to do that. But it does make me nervous in that user talk pages are often used to raise conduct issues. I hope you don't have any concerns about my conduct? As to whether the Guardian article is all puffery, there does seem to be a grain of truth -- at least it's true that Giuliani has been working with Chapman College and his findings have been presented at conferences. Also, the Chapman group is confirming a correlation between Giuliani's radon observations and the L'Aquila EQ. I'd been waiting for your comments but I hope you'll agree that those facts deserve a mention in the article. I don't know the status of the 'worldwide early warning system', whether it's just a glimmer in Ouzonov's eye, or whether there's anything more to it than that. I don't know anything about Pullinets, but if he's working together with Freund and Ouzounov, with theories about EM signals in the atmosphere, maybe the whole group deserves some sort of mention in the article? Of course they're 'fringe', but so is the entire topic -- at least in the sense that mainstream seismologists think that these prediction methods are never going to work at a functionally useful level. JerryRussell (talk) 23:57, 1 September 2016 (UTC)

One other topic I had raised in this rather wide-ranging discussion, is whether anything needs to be said in this section about the manslaughter convictions. I've since realized that the topic has at least been covered pretty well elsewhere at Wiki, and I'm feeling you might be right that it doesn't belong in this section. Maybe a little more could be said up in the 'evaluating predictions' section, where it's already got some coverage.

My ratio of talk page conversation vs. actual edits to the article is pretty high. Maybe I should just be bold, and not talk so much? JerryRussell (talk) 00:03, 2 September 2016 (UTC)

I didn't mean to suggest any "conduct" issues with you, unless perhaps that we both do tend wander quite widely. (Although I wonder: when you commented at 20:11 4 July about discussing "conduct issues only on user talk pages", were you perhaps thinking that user talk pages are only for conduct issues? Not so!) While here is the proper place to discuss what should go (or not) into the article, or how to assess various material and sources, discussion of our own particular views and judgments might be more appropriately on a user talk page.

A particular criticism I have here is that you (like most other editors) seem to be approaching the topic from the bottom. That is, from the perspective of a lot of interesting details. The problem is, there are a LOT of interesting details - I estimate at roughly two, perhaps even three, magnitudes more than you likely realize - and focusing on what is closest to hand (eye?) misses the overall picture, and quickly leads to imbalance. I think a better approach is top-down: consult reviews of the topic by the top-rated experts to determine what sub-topics (breadth) the article should cover; then, for a desired article length, dig down into the details only so far as does not create an imbalance in the breadth of coverage. E.g., to cover a forest fire we would not look at individual trees burning.

In that regard, sorry, no, these "facts" you refer to do not deserve a mention in the article. What's notable about Giuliani is his single alleged prediction (ballyhooed by the mass media into something more than a prediction), not his method, of which we know nothing other than it is based on "observation" of radon levels. Likewise, the manslaughter convictions are about a hazard in making, or failing to make, predictions, which has already been included. But even if that had been included elsewhere, it has nothing to say about the success or failure of Giuliani's "prediction".

I'm running out of time, I'll just ad that, regarding "bold" editing: lots of talk is how we avoid big bumps. ~ J. Johnson (JJ) (talk) 00:12, 3 September 2016 (UTC)

I think a better approach is top-down: consult reviews of the topic by the top-rated experts to determine what sub-topics (breadth) the article should cover; -- sounds like a good idea. Any reviews you'd like to recommend?

More on Giuliani at Italian Wikipedia, confirming that he spoke at the AGU conference in San Francisco: 2

Pullinet's team looks formidable: 1 JerryRussell (talk) 04:51, 3 September 2016 (UTC)

I think Geller's 1996 review in GJI was the most comprehensive, and best, review yet seen. The ICEF review of 2011 – requested by the Italian government after L'Aquila and headed by an international panel of experts – is also good, and still reasonably up-to-date. Zechar (2008; full citations in the article), does a good review in the first chapter of his dissertation. Kanamori (2003) is also good. Undoubtedly others that I don't recall off-hand. ~ J. Johnson (JJ) (talk) 23:26, 3 September 2016 (UTC)

VAN v. truth check

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

As VAN authors/promoters are trying to hijack the article, why not look at real science behind seismicity: http://seismo.info, with science behind it and the obvious proof. Fascinating stuff. 31.185.124.108 (talk) 16:08, 25 August 2016 (UTC)

Hi 31.185.124.108, I hope you don't view me as a VAN promoter? As to IP202, I think he is, but we've done our best to utilize his expertise without allowing him to 'hijack'. If you have concerns about the neutrality of the article, could you be more specific, rather than casting aspersions?

As to the Omerbashich material, I agree it's fascinating, but unfortunately we can't use it for the Wiki mainspace article. There has been no coverage of Omerbashich in mainstream reputable sources, and all the materials you link to are self-published. Have you seen suspiciousobservers.org? Davidson does have at least one peer reviewed article, and I think his views are broadly similar to Omerbashich.

Mainstream is hardly associated with reputation let alone truth. So much about the article's neutrality (as you imagine it). 31.185.124.108 (talk) 19:10, 25 August 2016 (UTC)

31.185.124.108, welcome to Wikipedia. Terms such as 'neutrality' have specific technical definitions here. We hope that the articles have some sort of relationship to 'truth', but the main goal is 'verifiability' as befits an encyclopedia. A good place to start if you aren't familiar with the rules here is Wikipedia:Five pillars. Do you know if any aspect of Omerbaschich's theories can be verified in mainstream sources or peer reviewed journals? JerryRussell (talk) 19:33, 25 August 2016 (UTC)

have you seen this discussion? Talk:Nobel_Prize_controversies#Omerbashich_claim The IP editor involved in that controversy was blocked for violations of Wiki policy. Owing to WP:DUCK you are going to have to be on your best behavior, otherwise the same thing is likely to happen to you. I'm sorry to have to bring this up here, I would normally want to use your talk page, but since you're an IP editor I have no way to know if that page will be stably associated with you. If your goal is to get materials into Wiki article space, you'll have to work within the policies. JerryRussell (talk) 20:14, 25 August 2016 (UTC)

Good catch, Jerry. You might also note the #Dr. Omerbashich earthquake prediction section above, where we have previously discussed http://sesmo.info//. That section was instigated by 91.203.111.4 (now blocked), who has also joined in at Talk:Nobel_Prize_controversies#Omerbashich_claim. If the content, tone, etc, aren't suspicious enough, check the geolocation: 31.185.124.108 comes from the same area (Mostar/Sarajevo). WP:DUCK is exactly appropriate. ~ J. Johnson (JJ) (talk) 23:28, 25 August 2016 (UTC)

As this topic was opened by an anonymous IP spamming some previously discredited nonsense, and has little prognosis for improving the article, perhaps some passer-by would kindly close it. ~ J. Johnson (JJ) (talk) 22:19, 3 September 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Is the entire subject of earthquake prediction fringe?

As several editors have commented that "the entire topic" of earthquake prediction is fringe, it seems necessary to address this. I suspect this notion arises from not fully understanding the difference between the idea, view, position, or claim that prediction of earthquakes is possible (either generally, or by particular methods), and the scientific study of whether such prediction is possible.

As an example: claims that animals anticipate earthquakes are rejected by most scientists, and therefore are "fringe", per WP:FRINGE. But studies of whether animals have anticipated earthquakes - done scientifically, and not, say, by post-event interviews of pet owners - are generally accepted as scientific contributions, and therefore are not fringe.

Same thing for "earthquake prediction" as a whole. Whether earthquakes might have any precursory or recurring characteristics is a proper and acceptable topic of scientific study. It is also notable, due to great societal interest (even by lay-persons), and having accumulated a large literature. To summarize plain and simple: earthquake prediction, as a subject or field of scientific study, is not fringe. And it is notable.

The view that prediction of earthquakes 1) has not been successfully demonstrated, and 2) may even be inherently impossible (repugnant as this seems to many WP editors, and even to a few scientists who claim they have predicted earthquakes), is not fringe either, because it is "broadly supported by scholarship in its field". As amply cited in the article itself. (Although likely less clear than before, due to accumulated deletions and alterations.) ~ J. Johnson (JJ) (talk) 22:39, 3 September 2016 (UTC)

Hi JJ, did you notice that the RfC was closed with the finding that "practically the entire subject of the article is fringe"? And I think that's true, if the subject of the article is 'useful, successful and proven earthquake prediction methods.' Because none of the proposed methods have been demonstrated to the satisfaction of broadly supported scholarship, right? Also, "The case was not made that policy requires all Fringe to be removed on sight."

If you think the article is less clear than it could be, are there any changes you'd like to make? I know you want to get rid of 'natural time', but again the RFC close said it should be included on an as needed basis.

As to the formulation that EQ prediction "may" be inherently impossible, what about that weasel word 'may'? It's hard to disagree with a 'maybe', but do you really think there's a broad consensus that EQ prediction *IS* inherently impossible? I'm guessing that such an extreme formulation is probably a fringe view, hardly anybody would come out and say that except Geller. But, I haven't read enough in this field to know. JerryRussell (talk) 23:04, 3 September 2016 (UTC) tweaked JerryRussell (talk) 23:33, 3 September 2016 (UTC)

Of course I saw the RfC summary, and particularly Geogene's comment that "practically the entire subject of the article is Fringe." That was out of place, as it was not a subject of the RfC, and I would say it is wrong. But closing such an extended discussion is such a thankless job (and good help hard to find) that I didn't want to belabor Geogene about it. That you or anyone else should think this subject is fringe arises (I believe, and as I have said) from not understanding the distinction between 1) a view that earthquake prediction is possible (either generally, or by some specific method), and 2) the scientific study of whether such prediction is possible.

If (as you properly qualify) the subject is about "useful, successful and proven earthquake prediction methods", there is no article, because there are no such cases (impassioned advocacy notwithstanding); that is an empty set, on account of those three qualifying words. Nor is the subject just about methods (successful or not), as without knowing something about the nature and challenges of prediction generally the reader cannot assess the credibility of either methods or specific claims of prediction.

I had thought the article was clear enough, except where material has been removed. What concerns me is where any of this is not to clear to you, and that is something you have to show me.

I note that, in science, "may" is not a "weasel word" as that phrase is perjoratively used on Wikipedia. It is a frank and express acknowledgement that we don't know all things, and there is always a chance we are wrong. (Although many self-styled "skeptics" really should give due respect to how small that chance often is.) On that basis I doubt that there is broad consensus that prediction "*IS*" – absolutely, in all senses and formulations, without caveat or qualification – "inherently impossible". There is a lot of argument (see the sources) that prediction is not merely impossible, but inherently so (meaning that the difficulty is not just a simple matter of inadequate funding). But lacking absolute proof of that, or even knowledge of all future developments, the assertion needs to be qualified with the "may". It is the exact converse of what one scientist said about prediction: it "may be possible in the future". (If his funding isn't cut?) Or: it may be impossible.

As to "hardly anybody would come out and say that" – well, Geller was joined by David Jackson, Yan Kagan, and Francesco Mulargia in stating unambiguously that "Earthquakes Cannot Be Predicted". And other voices in the Nature Debates. They are certainly not fringe, even if many other authors are still hopeful. ~ J. Johnson (JJ) (talk) 23:18, 5 September 2016 (UTC)

Prediction and forecasting, two sides of the same coin?

I've been reading thru the ICEF report as JJ suggested, which was excellent advice to get a broader view of the field. I found this definition on p. 327 for the difference between prediction and forecasting.

Predictions and forecasts both make statements about future earthquake activity based on information available at the time; that is, they provide prospective (before-the-fact) rather than retrospective (after-the-fact) earthquake information. In this report, the Commission distinguishes between a prediction and a forecast using a strict dichotomy. A prediction involves casting an alarm — an assertion that one or more target ruptures will occur in a specified subdomain of space (subregion) and future time (subinterval). Predictions are therefore prospective deterministic statements: if a target event occurs in the alarm subdomain, the prediction is a true alarm; otherwise it is a false alarm (or type-I error). If a target event occurs in a subdomain without an alarm, the error is a failure-to-predict (or type-II error). A prediction can also be cast as an anti-alarm, a deterministic statement that no target rupture will occur in a subdomain [33].

Forecasts are prospective probabilistic statements: they specify the probabilities that target events will occur in space- time subdomains. The probability in a particular subdomain is a number P that ranges between 0 (no chance of a target event) and 1 (certainty of a target event).

Based on this, isn't it a trivial matter to convert any prediction into a forecast, and vice versa? Beginning with a prediction, simply add an estimate of the expected success rate as a probability, and voila, it's a forecast. And if you've got a forecast, just set a threshold of probability of an EQ above which you will issue an alarm, and you've got a prediction.

Personally, it's an easy choice. I want to be in the forecasting business. JerryRussell (talk) 01:06, 6 September 2016 (UTC)

In his 1st article in the Nature debate series, Geller says:

The public, media, and government regard an 'earthquake prediction' as an alarm of an imminent large earthquake, with enough accuracy and reliability to take measures such as the evacuation of cities. 'Prediction' is used exclusively in the above sense here...

That makes sense too, but it's a different definition from the ICEF above. It's not a clean demarcation, unless you specify the parameters for "enough accuracy and reliability" that would make such predictions useful for the defined purpose. JerryRussell (talk) 02:27, 6 September 2016 (UTC)

Exactly, in that it's not a "clean" (exact!) demarcation. But Geller's formulation (of the general conception of EQ prediction) is not different from the ICEF, only more detailed. I.e., the ICEF says "[a] prediction involves casting an alarm". Geller constrains the scope of such alarms as involving earthquakes large enough and imminent enough to justify evacuation of cities. Which (as seen at L'Aquila) can be quite a dilemma for the public officials. In an earlier version of this article Geller is quoted as to what would happen in Japan if 'anomalous data' are recorded:

an ‘Earthquake Assessment Committee’ (EAC) will be convened within two hours. Within 30 minutes the EAC must make a black (alarm) or white (no alarm) recommendation. The former would cause the Prime Minister to issue the alarm, which would shut down all expressways, bullet trains, schools, factories, etc., in an area covering seven prefectures. Tokyo would also be effectively shut down.

The Dilemma: To Alarm? or Not to Alarm?

At a cost of billions of dollars a day, and possible panic, so which side of the "To Alarm? or Not to Alarm?" dilemma (see chart) do you want to err on? What if some lab technician says "sorry about the last two times, but this time I'm really, really sure you should go out on a limb again"?

And how do you respond to a forecast that (and let's assume on very certain information) there will be an M 8 earthquake within the next two years? Would you evacuate Tokyo - or even just Portland and its suburbs - for the entire two years? There is no particular day when the probability of that quake will exceed the probability of it happening on any other day. There is no threshold that you can set ("The Bar" in the chart) that isn't either too high, and fails to predict (Type II error), or so low that it predicts ("alarms") for every day (Type I error). So, sorry, no, it is not a trivial matter to convert. ~ J. Johnson (JJ) (talk) 23:30, 6 September 2016 (UTC)

Hi JJ, I agree with everything you're saying above, and I think I'm getting a better understanding of your goals here.       [Great! Hang in there. -JJ]

I have a suggestion. What if we set things up so that searches and links for "earthquake prediction" and/or "earthquake forecasting" all redirect to the same page, which is entitled "earthquake prediction and forecasting". The lede of that article would give Geller's definition of a prediction, which is an EQ alarm that is sure and imminent enough to justify evacuation of cities. Then we would explain that the consensus of seismologists is that earthquake prediction to that level of confidence is currently beyond the state of the art. Then we would give a link to an article entitled "Earthquake Prediction Research". The remainder of the article 'EQ Prediction & Forecasting" would actually be entirely about mainstream views of forecasting methods that are generally thought to be promising or worthy of research.

The article "Earthquake Prediction Research" would basically be a WP:FRINGE POV fork WP:SPINOFF article, where all the failed precursor methods would be described, along with each method's ever-optimistic research advocates. Of course this article would also prominently give mainstream criticisms of all the methods. I think the organization of that article would be cleaner if each of the 'notable EQ prediction' sections was combined with the associated method that made each prediction.

The goal is for people to arrive first at the "EQ Prediction and Forecasting" article, which presents the mainstream POV. Then if they really want to read about prediction methods, having been told they don't work, they can go on to the fringe article.

The way the article is now, looks very WP:UNDUE because there's so much material on fringe methods, and each of them needs to be treated respectfully according to neutral voice, since they aren't so bad as to get WP:PS treatment. And the reader needs to be very alert to pick up on the transition from the fringe prediction methods, to the mainstream forecasting methods, and then back to the list of fringe predictions.

I have an additional comment about your question how do you respond to a forecast that (and let's assume on very certain information) there will be an M 8 earthquake within the next two years? but I'm not sure how relevant it is to the article content & structure, so I'm going to post it to your talk page. JerryRussell (talk) 03:41, 7 September 2016 (UTC)

A major obstacle to such an apprpoach is that POV forks are considered "inconsistent with Wikipedia policies" (see WP:POVFORK). Another major problem is how to have readers "arrive first" at your proposed summary/disambiguation article. If someone searches for an article "X", and article "X" exists, I don't know how to redirect them to "X and Y" instead. Your suggestion would have to be an "X and Y" article that replaces both "X" and "Y". And that would mean an even larger article tryng to cover a broader scope.

The best approach - well known in all scholarship - for closely related articles is that each succinctly distinguishes at the very start exactly what it covers, and doesn't cover (i.e., its scope), and references the other article(s). A separate disambiguation article should not be necessary.

In that regard this article's current lead could be better (and in my opinion, was), but has suffered from gratuitous editing. As one improvement I favor restoring Kagan's quote that prediction is particularly "the determination of parameters for the next strong earthquake to occur in a region." That is the public's general conception of prediction, for all "next" that are soon enough and strong enough to warrant an alarm.

The sub-sections under "Notable predictions" do not cover the methods or basis of the prediction (except in the most summary form) because they are about the predictions themselves, not prediction methods. The latter are covered generally in the preceeding section (Earthquake prediction#Prediction methods). Note that some of these predictions do not have well-defined methods (e.g., Haicheng). And there are some 26 cases, covering a variety of methods, by which the Loma Prieta quake was "predicted" (or forecast) in some sense.

However, underlying all of this it appears there is still deeper misapprehension here regarding "fringe", and particularly in connection with due and undue weight. Even if every method of possible earthquake prediction had been rejected by the mainstream community, there has been enough work that an article about those methods would not be "undue", and certainly not "fringe". Note also that "neutral voice" does not mean equal balance (as the Athenian editors want for VAN); that would be a falsebalance. ~ J. Johnson (JJ) (talk) 22:14, 7 September 2016 (UTC)

OK, Immediately after the policy section defining bad types of POV forks, there's another section about good forks. I didn't realize that other section has its own link and name, WP:SPINOFF, see also WP:SPINOUT. There's a right way and a wrong way to do this. So, next time I'm advocating a spinoff or spinout, I won't call it a POV fork. My bad.

If we have an article "Earthquake Prediction" that just re-directs to "EQ Prediction and Forecasting", that's the path that a search for EQ prediction would follow. The article "Earthquake Prediction Research" (or, if you prefer, "Earthquake Prediction Methods" or "Earthquake Precursors") would be the second result on the search, or maybe somewhere down the page.

I'm not saying my suggestion is the only possible approach, but especially if you & Elriana are working on a forecasting article, I hope we can consider it. JerryRussell (talk) 22:35, 7 September 2016 (UTC)

If I may assist here, there is Template:About used on 101,000+ pages:

This article is about earthquake forecasts accompanied with a warning. For earthquake forecasts without warning, see earthquake forecasting.

The statements above are not accurate, I guess, but the template could add to argumenting.--46.198.213.62 (talk) 22:53, 7 September 2016 (UTC)

In the first place, this article is not "about ... forecasting" (with or without "warning"), it's about prediction. A distinction which I had thought was adequately explained in the lede, and certainly should be clear after all this discussion, so I doubt that any single sentence can suffice to make the necessary distinction. I expect that when the complementary article is ready both will have "See also" cross-links, with the distinction explained in the lede.

Jerry: "an article ... that just re-directs" is not an "article", it is a REDIRECT, which (unlike any kind of article) has no content. So if "Earthquake prediction" and "Earthquake forecasting" are both REDIRECTS to "Earthquake prediction and forecasting", they both have NO CONTENT. *All* of the content of both articles would have to go into the combined article. And I have not heard any complaints that this article is so short it needs to be expanded to include another topic. Which, by intimate association, will then further blur a necessary distinction. ~ J. Johnson (JJ) (talk) 05:51, 9 September 2016 (UTC)

JJ, I apologize that I used incorrect Wiki terminology for redirects, as explained at WP:R. A redirect is considered a 'page' but not an 'article'. I was meaning to talk about redirect pages with no content.

Right now, we have an "Earthquake forecasting" page which is simply a redirect to this article, and has no content. In my opinion, some of the precursor methods may have some merit for short-term forecasting, even if they have failed as prediction methods.

So just to be clear, my proposal is:

1) Change the name of our existing article to 'Earthquake Prediction and Forecasting'.

2) "Earthquake Prediction" to become a redirect page pointing to this article.

3) From this article, per WP:SPINOFF, we create a new article called "Earthquake Precursors". This is less confusing than calling it "EQ Prediction", even though it would be covering all the methods that purport to be EQ prediction methods. According to the mainstream, these methods are generally deprecated for alarm purposes. By spinning off this material, I believe we would be able to overcome the existing article's problems with undue weight on the fringe idea that these precursors are actually useful as predictors.

I agree we could also go ahead with separate articles for 'prediction' and 'forecasting', and that see-also cross-links and templates would be a reasonable alternative to solve the problems we've been working with. But I feel pretty strongly that apart from specialists, almost everyone would consider "EQ forecasting" and "EQ prediction" as synonymous terms, and that my proposal is a good way to handle this confusion. JerryRussell (talk) 18:46, 9 September 2016 (UTC)

Jerry, I can envision alternate arrangments, but your proposal is not well thought-out. In the first place, the scope of this article is earthquake prediction, not "prediction and forecasting". You could argue for a change of scope, but (as I have been at some trouble to explain, here and elsewhere) there is good reason for distinguishing these terms. And if these topics are merged the readers would be confused by the explanations that would have to be added, and the article would become quite bloated.

Your notion of spinning-off "all the methods that purport to be EQ prediction methods" is confused because you are still hung-up on supposed issues of undue weight and fringe. I particularly reject your insinuation that the article has "problems with undue weight on the fringe idea that these precursors are actually useful as predictors." I have tried to make the article quite clear that the mainstream view is that all these precursors are not "actually useful as predictors". (And taken flack for not being "neutral".) If something contrary is been slipped in point it out and we can remove it.

A further problem with "Earthquake precursors" is that it is only half of one section. (Please review that section.) Another problem is that you have not considered the other sections of the article and how they relate. And while you are proposing an "earthquake prediction and forecasting" article, you are also agreeing to "separate articles for 'prediction' and 'forecasting'". So it is quite unclear what you want. I suggest that any proposal on how this topic area might be covered with various articles really should have an outline of the particular topics each article would have. ~ J. Johnson (JJ) (talk) 22:39, 10 September 2016 (UTC)

JJ, thanks very much for your remarks, and your consideration. I had been hoping that other editors might join the conversation. Since my proposal hasn't attracted much interest or attention apart from your review, I'm feeling that I should withdraw it at this point. Perhaps when the article on 'forecasting' is completed, and the various sub-topics are sorted out, the issues I've been concerned about will be mitigated without any need for the scheme I suggested. JerryRussell (talk) 04:10, 13 September 2016 (UTC)

I am quite pleased to be working with someone who can accept criticism without construing it into any kind of personal attack.

I suspect your concerns arise largely from not understanding certain aspects of the material (or, heaven forbid, its presentation!!); expressing your concerns has been useful in assessing where other readers may have similar difficulties. I am considering how that may be addressed, and have made some adjustments to the lede. And yes, I think having an article on forecasting would provide the contrast to better distinguish these topics. ~ J. Johnson (JJ) (talk) 23:16, 13 September 2016 (UTC)

Contingency table

46.198.213.62 (IP202?) suggested a new version of the contingency table for successful & unsuccessful earthquake predictions. The new table presents a second alternative alarm type: a 'state alarm' as opposed to a 'public alarm'. I'm not sure I understand what this means. It sounds like maybe the state is trying to notify only specific parties? If so, I see enormous liability as well as information security issues.

But perhaps more importantly, all material needs to be sourced. Otherwise it's Original Research. The pre-existing contingency table doesn't give a source either, but it seems to follow the sourced text very closely, so it seems OK to me. I would say that unless IP202 can provide a reference, the new table needs to go. JerryRussell (talk) 05:41, 14 September 2016 (UTC)

"State alarm" means giving alarm to ministry head only. We can remove it until sourced. I would very much like to see an expert's opinion (and maybe sources) on this table, though.--IP202-46.198.213.62 (talk) 07:19, 14 September 2016 (UTC)

	If Quake		If No Quake
	Cost in Lives	Economic cost	Cost in Lives	Economic cost
Public Alarm	min due to EQ max due to panic	max, to be on guard economic disruption insurances until EQ EQ losses	max due to panic	max, to be on guard economic disruption
Government Alarm	moderate due to EQ	some, to be on guard EQ losses		some, to be on guard
No Alarm	max due to EQ	EQ losses

I removed the new table for now. If the alarm is given to the ministry head, what does the ministry head do with the information?

One of the sources for the paragraph next to the contingency table was Mason 2003. That reference had disappeared from the source list, but I found & fixed it. It's a very mathematical treatment of statistical evaluation of error rates for contingency tables, and has little or nothing to say about social costs of errors. Geller 1997 and Atwood & Major 1998 are the other sources given in the associated text related to the existing contingency table. JerryRussell (talk) 14:21, 14 September 2016 (UTC)

Jerry: A small detail of bibliographical practice: because "Mc" is the elided form of "Mac", and often used interchageably, "Mc" is generally collated as "Mac". Thus "Matthew" really does come after all the "Mc" entries. ~ J. Johnson (JJ) (talk) 20:36, 14 September 2016 (UTC)

JJ, I don't have a strong preference about this. But, Wiki's article on the topic Mac and Mc together agrees that sorting Mc and Mac together is a longstanding practice as you say, but also that ISO 999 and modern style guides such as Chicago Manual of Style recommend separating Mc and Mac so that names will be sorted as they are spelled. JerryRussell (talk) 15:20, 16 September 2016 (UTC)

Headships take measures in the form of readiness exercises. The driver of the ambulance or a rescue group coordinator think they have a routine readiness test. Practically only a few top people know about it.--IP202-46.198.213.62 (talk) 14:59, 14 September 2016 (UTC)

"Headship"?? What an intriguing term. Googling on it finds:

• 10 Ways to Exercise Christlike Headship
• What Does It Mean to Be the Head of the Home? - Grace to You
• Headship - The Marriage Bed - Front Page
• Headship | Define Headship at Dictionary.com
• Headship of the Husband in the Bible
• What Does the Bible Say About Headship?
• 5 Myths of Male Headship
• I Believe in Male Headship

For "headship alarm" we get:

• Extreme Headship: A Case Study in Educational Leadership and School ...
• Making your Way in Headship
• Race, female headship, and delinquency: A longitudinal analysis ...
• The seven deadly sins of executive headship

And so forth. In the context of alarms this term "headship" is quite novel, even unknown. As IP202 didn't bother with any sourcing it does appear that he is just making it up. "Original", yes, but I wouldn't deign to connect this with "Research".

"State alarm" (meaning, I presume, "state" as in government, and not "state of alarm") is also novel. But it fits right in with Varotsos' mode of issuing "predictions" only to some government agency, and letting the government (Minister) have the burden of issuing, or not issuing, a public alarm. That way, if something happens, they can always say "we told you!" If not, just let it pass. Or complain that a "prediction" was not supposed to made public. ~ J. Johnson (JJ) (talk) 20:33, 14 September 2016 (UTC)

In Greece this model seems to work just fine. In Japan the losses are so great, either way, that they cut funding. In Greece this is not the case, there is balance and there is flow. In rare cases like 2008, there are public alarms.--IP202-46.198.213.62 (talk) 06:23, 15 September 2016 (UTC)

Huh? You are talking nonsense. Your table does not illustrate any kind of model, it's not even complete, and the terms are undefined, amounting to gibberish. The only "model" it even suggests is to distinguish between "public" and "government" alarms, a distinction heretofore unknown in the literature, and serving (in Greece) only to give VAN more bases for prevarication.

You "would very much like to see" some sources on your table? Wow, so would we. How about where you found the table? Oh, it appears you just made it up. (I won't ask where you got it, as I think we all have a pretty good idea.) Well, the absolute first principle at Wikipedia is WP:Verifiability. No sources? It doesn't belong here. And you are just wasting our time. ~ J. Johnson (JJ) (talk) 23:12, 15 September 2016 (UTC)

The new table clearly demonstrates that JJ's table is Original Research. Warning systems must exist within the national emergency management framework although that framework may work against effective local warning (last conclusion). There is more ↑↑ to study, in order to avoid making a wrong impression on alarming via the article.--IP202-46.198.213.62 (talk) 05:54, 16 September 2016 (UTC)

In Greece this model seems to work just fine. IP202, this is a fascinating claim. Please clarify how you know this?If you can't provide a source, are you stating that you have inside information that isn't known to the Greek public?

Your sources about sociology of alarm systems look very interesting. I just skimmed the abstracts, I haven't read the full texts yet. I don't agree that our existing article & contingency table are "Original Research". But they may represent an outdated or incomplete representation of the current state of the art, in light of these papers you've cited. JerryRussell (talk) 14:40, 16 September 2016 (UTC)

Yes, a fascinating claim. Even more fascinating (I think) is IP202's assertion that: "The new table clearly demonstrates that JJ's table is Original Research." That is completely bull fart. Which is to say, offends the nose, but has absolutely no substance. He made something up, something that appears to be quite novel in the field of forecasting, with entirely unknown (and undefined) terms, and zero sources. When called on that, he promptly turns around accuses me of Original Research. (A pattern we've seen before, where he is good at echoing what anyone else says.)

Now he throws some links at us, but I don't see they have anything to do with his table. (Which is still crap.) His references seem to be more on the sociological challenge of adequately commmunicating an alarm. IP202 seems to have totally missed that the point of the "Dilemma" table presented in the preceding section (and in the article) is the dilemmma of whether to issue an alarm, or not.

The topic of this section (as opened by Jerry) is IP202's totally unacceptable table. For IP202 to then challenge the "Dilemma" table is off-topic, as that has nothing to do with IP202's table, and is just obfuscation. If he thinks there is a problem with the existing table he is, of course, free to raise it, but that is a separate topic. ~ J. Johnson (JJ) (talk) 22:45, 16 September 2016 (UTC)

Existing contingency table

JJ, as you suggested I'm opening a new section for this discussion.

IP202's third reference, ↑, "Communication of Emergency Public Warnings", Mileti & Sorensen 1990, section 3.1.3, asserts that "the public simply does not panic in response to warnings of impending disasters", except in very narrow circumstances that would rarely apply to EQ predictions. Also it states that the "cry wolf" syndrome is greatly overrated, and that people can understand that false alarms are part of the game. It says that people prefer to have access to multiple sources of information, if possible. All of this seems to be relevant to the text & contingency table in the "Evaluating EQ Predictions" section. JerryRussell (talk) 00:27, 17 September 2016 (UTC)

Jerry, what is your point? Is there any particular problem you have found with the "existing contingency table"? (And you are referring to the "Dilemma" table, corrcect?)

You cite IP202's reference that "the public simply does not panic in response to warnings of impending disasters", even though several ocassions of panic are noted in the article. You cite the same source that the "cry wolf" syndrome is overrated, even though that is not mentioned in the current text, and certainly not in the table. But, keeping in mind it was only a month or so ago that IP202 disparaged papers from 1996 as "two decades old" and therefore no longer applicable, I point out that your source is from 1990. More relevantly, also supplanted by later sources, such as Atwood & Major (1998), who said there is a "cry wolf" effect. But so what? What has that to do with the table? ~ J. Johnson (JJ) (talk) 21:01, 17 September 2016 (UTC)

JJ, I agree that IP202's source is from 1990. As you've made the point with IP202, earlier sources are not automatically supplanted by later sources such as your 1998 source. Optimally, it would be a good idea to carefully read both sources, as well as to do research to find out if there has been more relevant work in the field since 1998.

When you think about it, 'panic' is a rather subjective term. I hope you're not "in a panic" about our dialog here, but someone might (hyperbolically) say so. Are there any objective measures of public panic? How many fatalities have been recorded as a result of public panic over false-alarm EQ predictions? For that matter, how many false alarms does it take before people start ignoring EQ predictions, and what can officials say to mitigate that problem? I don't know, but it does seem relevant to our article.

I'm not asking for any immediate action regarding the article or the 'Dilemma' contingency table. We're all volunteers here, and I'm not sure this even deserves to be high priority on our list of action items. But just by way of being cordial and conversational, I wanted to express my views about IP202's issue. JerryRussell (talk) 20:34, 18 September 2016 (UTC)

As I asked before: what is your point? And: how can you have any "views about IP202's issue" when (as far as I can see) IP202 has not raised any issue?

So you want to ("optimally") "carefully read both sources, as well as to do research to find out if there has been more relevant work in the field since 1998." Why? Because you think "panic" is too subjective? Hey, take a look around: ALL terms are subjective, when you take a really close look at them. Perhaps you want we should just put this article on hold until all of the words used in it can be re-defined to remove every last bit of subjectivity? That would be absurd. As it is, I believe every instance of the use of "panic" in the article (if that is your issue?), including the existing table, is based on suitable sources, and presents no more problem here than it does in the sources. This article could use more work, but trying to find "objective measures of public panic" where it is not at issue doesn't warrant any priority. It is an utter waste of time. If you have a question, fine, ask, but I don't see any issues here that warrant discussion. ~ J. Johnson (JJ) (talk) 23:47, 19 September 2016 (UTC)

JJ, there are lots of things in the article that are objective and numerical (or at least trying to be), such as earthquake amplitudes, prediction success rates, and fatality rates. But, it's OK with me if you don't want to spend time researching this. As I said, we're all volunteers here. JerryRussell (talk) 01:02, 20 September 2016 (UTC)

In Wikipedia we cannot even use pies with political parties on elections, because of original research. Why include one party and not the other? Have all the sources provided the same pie? Figures create impressions. So please find JJ' misleading argument map published somewhere, even 30 years ago, or remove it from the article.--IP202-46.198.213.62 (talk) 03:11, 20 September 2016 (UTC)

IP202, the policy doesn't say not to make any graphs. It says to avoid misleading content in graphs. There's nothing misleading about the graph, because JJ does have sources indicating that panic and 'cry wolf' syndrome are both problems with false alarms. I put a mention of Mileti & Sorensen 1990 into the article, but it can't supersede the later references JJ provided. So I agree with JJ: there's no need for a change to the contingency table, considering the information we have available. JerryRussell (talk) 19:09, 20 September 2016 (UTC)

Jerry, it's not that I don't want to spend any time on this, it's that this trifling, petty little tangential point on whether people actually panic due to earthquake alarms (or whatever your concern is about) is not worth spending any time on. By anyone.

IP202 calls the "Dilemma" table an "argument map", but in his typical **ck-headed fashion gets it all wrong. In the first place, it simply does not map the structure of an argument, it includes no elements of argument such as premises, conclusions, etc. Quite simply: it is NOT an "argument map". He also calls it "misleading", but (just to be pedantic) he does not assert that it is misleading. And he most certainly has not shown how it is misleading in any point, nor in any fashion. Until he can demonstrate that it is misleading, this discussion has no basis.

And what BOTH of you have missed is that whether the exact extent "of panic related to public disaster forecasts" has been over-estimated (or not) is so extremely tangential as to be irrelevant. It seems that BOTH of you do not understand the point developed in the text: the DILEMMA (is that large enough to notice?) presented in deciding whether to issue an alarm, or not. That is, lacking absolute knowledge there is a certain chance of a Type I error (false alarm), and a certain chance of a Type II error (failure to predict, or miss), and one can be reduced only increasing the other (see Mason). This is BASIC STATISTICS. It is a matter of the classic trying to steer between Scylla and Charybdis. (Read it.) Incidentally, if you quarter and then rotate the adjacent image 90° counter-clockwise the rock (Scylla) corresponds to the "No Alarm/Quake" quarter of the table, and the whirlpool (Charybdis) corresponds to "Alarm/No Quake". Would this image make the point any clearer?

The mentions in the table of "panic", as well as "cost of alarm", "economic disruption", and "great losses" are incidental to show the TRADE-OFFS in being too sensitive in issue an alarm, or too insensitive. "The Bar" corresponds to where one sets the the decision threshold. All of this is BASIC STATISTICS.

As Jerry said, there is no need to change the able. Not the least because IP202 has shown nothing wrong with it. ~ J. Johnson (JJ) (talk) 05:32, 21 September 2016 (UTC)

JJ, would you consider humoring us by changing the text in the table for the false alarm box to: "False alarm: cost of alarm, public distraction and reduced economic productivity. 'Cry Wolf' effect?" I can't easily make the change because I don't have the source file for the "Alarm_dilemma.png" graphic. JerryRussell (talk) 15:41, 21 September 2016 (UTC)

Not enough: timing is also left outside, evacuation decisions and false alarm costs have already been evaluted--IP202-77.69.68.33 (talk) 16:25, 21 September 2016 (UTC)

IP202, now you're missing JJ's point. He's trying to illustrate a basic statistical concept with the graphic. Yes, the costs of false alarms (and the benefits of valid alarms) can be effected by the timing of the alarm, the distribution of the information, what information is conveyed, and so forth. But no matter how you do it, the basic dilemma is the same. JerryRussell (talk) 16:32, 21 September 2016 (UTC)

Of course it is not the same. Because a scientist does not alarm the public, the state should decide on this. The scientist alarms the state only, or he gets prosecuted if he alarms the public. So what is the definition of the alarm here? State or public alarm? But if this is answered, well, the diagram info changes a lot.--IP202-77.69.68.33 (talk) 17:12, 21 September 2016 (UTC)

IP202, I don't understand how it changes the basic point of the diagram at all, whether the scientist issues the alarm directly, or whether a state minister does the job. Either way, there are costs for false alarms. If the state decides not to issue a public alarm, that minimizes costs, but also minimizes benefits of a correct prediction. Also, we need a reference that discusses this concept of a "state alarm", otherwise we can't even mention it in the article. JerryRussell (talk) 17:50, 21 September 2016 (UTC)

Sorry, I'm rather short of "humor" at this time, as well as pressed for time and attention. And (as I just said) this petty tangential point on whether people panic is not worth spending any time on. Nor has IP202 shown any basis for his insinuation that the table is "misleading". Even if I had "time enough, and humor" (to paraphrase one of my favorite poems) to attend to every little nit-picking objection IP202 might raise, he would keep raising them until we were singing paeans of admiration for VAN.

IP202 keeps throwing out all kinds of chaff (and what are we to make of his incomprehensible "timing is also left outside"?), but he seems incapable of making definite statements about what is bothering him. (Aside, that is, from anything that might cast the slightest doubt on VAN.) But no matter how much he quibbles, it is just as you said: the basic dilemma is the same. No matter how one sets the decision threshold, there is always a trade-off.

I point out that trying to follow these several discussions gets really hard when they get all intermingled. As the heading of this section is "Existing contingency table", I suggest that all discussion of "headship", and "state" versus "public" alarm be restricted to the preceding section (on IP202's "Contingency table"), where it was introduced and has already been discussed. ~ J. Johnson (JJ) (talk) 21:13, 21 September 2016 (UTC)

IP202 apparent COI?

It seems to me that with his most recent posts, IP202 has painted himself into a corner. He's already admitted that there are no sources available to back his claim that "state alarms" are being secretly given to "headships" in Greece, and I don't see how he could make that claim without having some sort of inside information. I think this creates an apparent WP:COI, the extent of it is hard to judge without a specific declaration on IP202's part.

In accordance with the outing policy, it's not proper to press IP202 for more information. But with this evidence of a COI, and with all the complaints from JJ about IP202's bias -- I feel I'd also like to ask that IP202 not make any edits directly to the page? Come here first, and let's discuss. That's what always happens anyhow. JerryRussell (talk) 20:42, 18 September 2016 (UTC)

VAN has been for all years submitting its predictions to the state, by telegrams, faxes, etc. The state has lowered the flow of funding to VAN, but has not cut it down, as Japan has. It is common sense.--IP202-46.198.213.62 (talk) 21:38, 18 September 2016 (UTC)

Hi IP202, the part about the government ministers sending out emergency crews on "routine drills" in response to VAN alerts, doesn't strike me as simple common sense. For all we know, the government might just throw the telegrams in the circular file. But I'll accept that this suggestion is just your conjecture. Thanks for your clarification. JerryRussell (talk) 22:16, 18 September 2016 (UTC)

I don't believe the anti-outing policy prevents us from pressing someone about who they are; it just prevents us from releasing such information until the user himself does it. (Which a user might do to show he does not have a COI.) But it is a fair conjecture (given his consistently biased and non-neutral POV, and his intimate familiarity with VAN's work), and without actually identifying him, that IP202 has a conflict of interest here. So I concur that it would be preferable that IP202 not edit the article directly. ~ J. Johnson (JJ) (talk) 23:51, 19 September 2016 (UTC)

Giving examples on taking measures by the government, is a visualization of what alerting only the government could mean, in response. We 've been living and talking on VAN for decades in Greece, so things are more than obvious for us here. The case of JJ's misleading argument map has nothing to offer to VAN method itself, rather than creating a false impression that affects the whole article. Which reminds me there are still unresolved issues on VAN like the two examples used that were admitted they were added only to support Geller's point of view and should have already been removed by now. This is my answer to the suggestion to talk first, and be patient, rather than editing the article directly.--IP202-46.198.213.62 (talk) 03:11, 20 September 2016 (UTC)

IP202, it's important that Geller's point of view regarding VAN is also represented. But -- regarding the two examples, are you talking about this text: "...such as (typically) a prediction of an earthquake at either 300 km "N.W" of Athens, or 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit.? I think you're right, you and I had agreed that selection of two examples from a list is OR evaluation of primary source material, and we should have removed it. I apologize for the delay. I'm removing it now. Hopefully JJ will agree too.

Regarding your point that the cost of false alarms can be mitigated by issuing the alarms at "headship" level only: are you sure this hasn't been discussed in the literature? It does seem very plausible. JerryRussell (talk) 18:25, 20 September 2016 (UTC)

Not plausible at all, actually. But the basis for inclusion not how plausible it seems to you, but WP:Verification. And you will note that I have already pointed out (two sections above, and it would be so nice if every could just stay on topic) googling shows no significant uses of "headship" in this context.

IP202: As there is NO "argument map" (as demonstrated in the preceeding section), and you have made no showing of how the table discussed in the preceding section is misleading in any way, I will request that you stop making such false and uncivil insinuations.

Jerry: if you had an article about (say) some temple, and you had a choice of a hundred images, would it be OR to select two of them? F*****G NO! As I recall (and why should I spend any more time responding to such asinine complaints when the complainer is being a jerk?): that VAN are so non-specific is a key complaint, which is soundly supported by the sources. But what you removed is not the assertion of non-specificity, but its illustration. Furthermore, it is absurdly specious of IP202 to complain of the selection of these telegrams as examples as they are the very examples selected and provided by VAN themselves. (I.e.: in Varotsos & Lazaridou (1991), and Varotsos, Lazaridou, Eftaxias & Antonopoulos (1996a).) That in their own words they demonstrate the very point the experts complain of is not Original Research. That IP202 is so sensitive to anything that illustrates a criticism of VAN shows (and hopefully brings this discussion back on topic) that he is too biased to edit here. It is a strong indication that he has a conflict of interest.

In summary, the text removed is NOT "OR", and its removal violates NPOV in favoring a particular view not supported by the mainstream. I expect it to be restored. ~ J. Johnson (JJ) (talk) 05:56, 21 September 2016 (UTC)

JJ, if you had an article about some temple of Kali, and you chose two images showing Kali eating babies (out of hundreds that might show the fine architecture of the temple, and many more images showing Kali doing good deeds) then of course there would be complaints about OR, NPOV, and so forth. I don't know if you follow any political articles, but I've seen major brouhahas about choice of candidate photos.

Varotsos & Lazaridou 1991 is hidden by a paywall, but Varotsos et al 1996a is at Google Books preview, and I couldn't find where the two examples you cited were singled out from among many others. I could easily have missed something, I suppose. I found an appendix describing three EQ predictions, with copies of long, rambling telegrams giving explicit yet vague predictions in a way almost defying summary or description.

The text clearly states that VAN has been criticized for lack of precision in their predictions, so I don't see any violation of neutrality here. The section is still too long (giving disproportionate attention to VAN as it is), so I think that removal of text is generally a good thing. JerryRussell (talk) 16:06, 21 September 2016 (UTC)

Jerry, you are implicitly assuming that the two telegrams explicitly cited are not just significantly different from the rest, but are especially defamatory. It's like you think I deliberately passed-by a hundred exact, well-documented predictions, and went straight to two telegrams showing VAN eating babies. That is unwarranted. I think you didn't just miss "something", you missed nearly the "all" of the matter. Particularly: ALL of the VAN telegrams (at least the ones VAN has deigned to publish) are ambiguous. That you don't have access to some of the materials does not excuse you from filling in any gaps in your knowledge with what you think might be there.

"[W]here the two examples [I] cited were singled out from among many others" would be in the sources (as cited). The sources (all VAN) selected what telegrams to publish, and it seems fair enough to assume (especially reading the context) they were trying to show how thorough they were. There is, absolutely, no original research in "selecting" those items as representative examples.

Please keep in mind that this section is about IP202's evident COI, and the only reason we are discussing the telegrams is because IP202 said that material was "added only to support Geller's point of view". (And then you took the bait and removed that material.) This is IP202's persistent theme, to reduce all criticism of VAN to "Geller's point of view", as if all this is just an inconsequential dispute between VAN on one side and "Geller" (and yours truly) on the other side. That persistent and one-sided theme, being contrary to the preponderance of expert sources, shows (yet again) IP202's deep-seated bias, from which a conflict of interest is reasonably inferred. ~ J. Johnson (JJ) (talk) 21:19, 21 September 2016 (UTC)

JJ, I'm not saying that you did anything improper to cite the two telegrams. I tried to verify whether VAN singled out the two telegrams, but I couldn't find it. I believe IP202 is claiming that the two examples are prejudicial against VAN, but I'm not saying I agree with that. Even if you're correct that the examples are typical and were selected by VAN themselves, I don't think the article is significantly harmed by removing them.

So far, IP202 seems to be respecting the request to talk things out here, rather than make edits directly to the page, even though he has not admitted any COI. Even WP:COI editors are allowed to participate at talk pages, and make requests for changes. JerryRussell (talk) 21:48, 21 September 2016 (UTC)

But you did, in fact, say (just above, on 18:25, 20 Sep.) that you and IP202 "had agreed that selection of two examples from a list is OR evaluation of primary source material", and thus improper. This "had agreed" apparently refers to your comment of 03:17, 15 Aug. in the #Libel on VAN 1983-1995 section (where it was actually off-topic):

The selection of two examples [examples] is intended, as it says, to be a typical representation of problems with VAN predictions, namely a lack of precision. The selection could be criticized as "Original Research" but then again it might be equivalent to an obvious application of simple arithmetic.

"But then again" it actually is NOT original research, as I just previously explained. (And the article is harmed by removal of that material, as it is clear and unequivocale illustration of a point the critics make, and its removal is yielding to editorial pressure from a VAN partisan.)

Jerry: note that IP202 does not have to make any direct edits when he has you to do his carrying and chopping.

Note also (getting back on topic) that IP202 is WP:NOT "here for the encyclopedia". And his single-topic focus, his campaign to promote and burnish VAN, while removing or otherwise muting anything that might tarnish VAN, has been quite successful, in that for the last six months practically the entire (and voluminous) discussion on this Talk page has been about VAN. Even in regard of the contingency table in the "Evaluating" section (where VAN is not even mentioned) his comment (just above, at 03:11, 20 Sep.) is that it "has nothing to offer to VAN method itself". As if this article and everything in it must orbit VAN. So while COI editors are often allowed to participate at talk pages, IP202's participation here has been entirely disruptive in miring us in a morass of VAN, VAN, VAN. ~ J. Johnson (JJ) (talk) 22:36, 22 September 2016 (UTC)

JJ, as per your ongoing requests to stay on topic in each section, I'm going to comment here on the IP202 conduct issue. I'll start a new section about the two examples.

I agree IP202 is clearly partisan, and his account is WP:SPA. As to the charge that I "do his carrying and chopping", I'm sure you've noticed I've reverted his edits a couple of times, most recently his new contingency table. So I don't think this is fair. Also, I started this thread by stating that his comments about "state alarms" in Greece seem to indicate inside knowledge. I think you agree it was a good time to ask the question.

If there are behavioral issues with an editor, I believe there's a protocol. First you go to that user's talk page, and discuss the concern. If that doesn't resolve the issue, the next step is to post at WP:AN and ask for community sanctions. I hope you won't do that, because I think his contributions here have been helpful. But whatever you do, I don't think it's useful to continually bring it up here on this talk page, where conduct issues should generally be off topic. JerryRussell (talk) 00:05, 23 September 2016 (UTC)

Jerry: of course you have done some good work, and I am very appreciative of that. But sometimes you are a bit quick to leap, and my comment was meant specifically regarding your "telegram" edit. I did not mean to impugn the entirety of your work!

Regarding IP202's "behavioral issues": dealing with personal attacks does indeed usually start with a discussion on that user's talk. Only here 1) he doesn't have talk page (his IP address changing everytime his ISP assigns him a new one), 2) he is already informed of my complaints in this regard, and 3) it hasn't made any difference in his behavior. ANI would be the next stop, but to what point? Topic banning? That's where it would have to go. Are you ready for that fight?

As to IP202's editing behavior: that is fully relevant here. E.g., while ordinarily it would be quite legitimate to ask for explication of a point, or for sources, and even to question whether a section has gotten too long, it is quite a different matter when an editor shows a pattern of demanding sources on one side only, then complains of length, with a view of "fixing" the length by removing material he is opposed to. His behavior is to the end of affecting the content, and the treatment of the content, to a particular and non-neutral end. It needs be mentioned here, even continually, lest we, or anyone else that just happens to pass by, gets sucked into this slow drift of biased edits. ~ J. Johnson (JJ) (talk) 23:27, 23 September 2016 (UTC)

"Are you ready for that fight?" It's basically your choice, and my guess is you would probably get a topic ban or range block if you ask for it. I'm hoping you won't do that. If you look around at my editing projects at Wiki, you'll see I'm often working on fringe topics. I try to do it in a way that conforms to policy, but I do appreciate IP202's expertise about VAN. JerryRussell (talk) 00:31, 24 September 2016 (UTC)

I agree that IP202 is likely in a position to give us details about VAN, but I do not see that his editing ("contributions" makes me gag) have been useful. He is at best tendentious, but more accurately disruptive. VERY disruptive. As to his motivation, COI of some sort seems the most charitable explanation. ~ J. Johnson (JJ) (talk) 22:19, 24 September 2016 (UTC)

I've been reading the pages on WP:SPA and WP:ADVOCACY and thinking about appropriate and inappropriate ways for editors such as IP202 to participate. I believe it's encouraged for IP202 to use his expertise to work with us to get to a neutral treatment of the VAN topic.

To the extent that IP202 is trying to get us to ignore policies in order to promote VAN, that's not appropriate. But in general, I think IP202 understands NPOV in principle, though of course there are also disagreements about implementation. And most of the disruption that occurs, is because other editors (and not just me!!) have felt that IP202 has raised some valid concerns. We probably generate 10x as much text as IP202 writes, debating over his issues; but perhaps that's because it's not an easy thing, to know how this article should be written. JerryRussell (talk) 18:36, 25 September 2016 (UTC)

Your fundamental premise is flawed: IP202 is not "work[ing] with us to get to a neutral treatment of the VAN topic." Whether he "understands NPOV in principle" is rather irrelevant, as his activity (and that of the first Athenian editor) have been persistently non NPOV. That some of his "concerns" are plausibly valid on their face and considered in isolation is to naively miss the ploy: the selective use of a general rule. E.g., to cite NPOV ("balance") for adding VAN-positive comments and sources, but to cite length for removing alleged "negative" materiel. It's like typing an "F", then a "U", then the key below the "F", then a key below the "U", and considered in isolation all those acts are quite innocent. Right? But seeing the result right in your face, especially multiple times, is not so innocent.

As to wordiness: I do believe in details, and careful and precise statements, which tends to involve more words than making wild, ill-stated, and unsupported insinuations. There is also the aspect of communicating complex, subtle, or just unfamiliar concepts with others that don't have the same background. ~ J. Johnson (JJ) (talk) 23:45, 25 September 2016 (UTC)

VAN sections too long?

Please have a look at June 6 version of VAN, compared to today's size, which has been built on details under the negative point of view on VAN. There was an ongoing talk on the talk page, admins were called, but JJ kept on building anti-VAN content without listening and by reverting all IP202 edits. If Jerry had not appeared, JJ's hill would be a mountain now.--IP202-77.69.68.33 (talk) 16:57, 21 September 2016 (UTC)

If we can get a consensus on the talk page, I would support migrating a lot of the VAN material (including portions of the 'notable predictions' sections) out to the VAN method article, leaving behind a neutral summary in this article. JerryRussell (talk) 19:39, 21 September 2016 (UTC)

No. At least as in don't run ahead of the discussion and start deleting stuff. That the VAN material is running long is because, first, IP202 (and his colleague) wanted to add comments to all of the VAN criticism, and then, second, demanded more support and sources. That he now complains of the length (ha! what he really doesn't like is the criticism) is not honest dealing. It just more of his whip-saw campaign to shape the article to the liking of VAN. That he is now complaining of the editor is because he has little basis for complaining of the content. Except that (oh, my) now it's too long. ~ J. Johnson (JJ) (talk) 21:32, 21 September 2016 (UTC)

OK, I won't delete or migrate any more material without consensus. JerryRussell (talk) 21:50, 21 September 2016 (UTC)

Good. And thank you.

As the question keeps coming back up I am going to repeat a comment from 04:17, 15 Aug.: the VAN section has gotten longer because a certain editor keeps challenging what was originally summarized, requiring fuller explanations and citations.

For anyone just tuning in please note that for the last several months this anonymous single-purpose editor we call "IP202" has been adding text to burnish VAN and mute any criticsm, has repeatedly challenged anything he deems to be "anti-VAN" and insisted on more sources, and then raises this issue that the VAN section is too long. He has repeatedly attacked me because I am the main obstacle in his attempt to make this article a hosanna for VAN. ~ J. Johnson (JJ) (talk) 23:06, 22 September 2016 (UTC)

JJ, what do you think? Regardless of whether the problem is entirely caused by IP202's influence, would you now agree that the section has gotten too long? JerryRussell (talk) 00:08, 23 September 2016 (UTC)

No, I won't agree. That's IP202's issue, and if the section has gotten too long it is entirely due to his efforts. That is not acting in good faith, and he is just playing us, and causing us to waste more time. ~ J. Johnson (JJ) (talk) 23:43, 23 September 2016 (UTC)

The length issue was an important aspect of my argument below in favor of eliminating the two examples. So now they're back. Meanwhile, I still believe the section is disproportionately long, and you've stated that the reason it is so long is due to IP202's influence. I would think nobody would be better aware than you, that pressure from IP202 is not a good reason for the section to be so long. Or if you believe that VAN really is so important that the section on VAN predictions should be three times longer than any other section, I'm not sure I understand why. JerryRussell (talk) 00:12, 25 September 2016 (UTC)

I do agree with the points raised concerning not keeping the biased examples in the section and keep it in a logical size.

JJ' s unusual demand that nothing from his writing could be deleted or shortened does not serve Wikipedia. He has systematically excluded, without consensus, all VAN published responces to the criticism presented in the article, and just added, without consensus, more detail of the criticism to make it harder to be handled. This is not fair. This extends in the electromagnetics section:

1. The reader is left with the impression that it is impossible for SES to travel such long distances, when VAN have published how conductivity of neighboring faults, exceeding 100-1000 times the conductivity of the surrounding medium, give the path for SES to travel hunderds of kilometers and reach the surface having an amplitude comparable to the signals measured and reported by VAN.
2. There is no mention of the criteria that clearly distinguish SES from other electric signals of no precursory nature and the reader is left believing, as stated, that VAN do not account for (i.e. identify and eliminate) possible sources of electromagnetic interference (EMI).
3. The "Taken as a whole" paragraph has been answered by a series of VAN publications, also deleted by JJ, and only a negative study is selected to be presented there, too.

We cannot know who is right and who is not and in what extent, but when detail is presentend in criticism and is left hanging out there as if there was no published explanation/answer to the specific accusations, this imprints. If JJ insists in being absolute and not willing to contribute in consensus, I guess tagging back VAN section (perhaps EM also) could save my time and energy, too.--IP202-176.92.153.150 (talk) 16:16, 25 September 2016 (UTC)

IP202, I don't agree that VAN published responses to criticism have been excluded. All the references to VAN responses are provided at least in footnotes. In the 'Electromagnetic variations' section, VAN gets the last word, with the Huang article reference. In the EQ prediction section, VAN also gets the last laugh, with the apparently successful 2008 prediction and Papadopolous' lame sputtering in response. At this point, if there's anything that isn't crystal clear in the article, I would say it's VAN's very high false alarm rate. 1987-1995, 94 predictions, 10 hits at best. That's an 89% false alarm rate. The numbers are there in the article, but we don't really make it leap out at the reader.

Since 2001, nobody has even assessed how high the false alarm rate is.

Tagging the article really doesn't solve anything.

I appreciate your pointing out the biased examples in the section. As I mentioned below, I've now examined the data in Varotsos et al. 1996a and agree that the selection is biased. I'd like to solve the problem by deleting the examples here, but since my long-term goal is to move some of this excess detail to the VAN method article, perhaps in the meanwhile the answer is to add more examples and more detail. JerryRussell (talk) 17:58, 25 September 2016 (UTC)

Jerry, are you paying attention? How many times do I have to explain this? As I just said (above), the VAN sections got longer because IP202 insists on adding VAN's responses to the criticism, and keeps challenging what was originally summarized, requiring fuller explanations and citations.

Now he wants to quibble. E.g., in his point #1 he says: "The reader is left with the impression that it is impossible for SES to travel such long distances....". And why shouldn't the reader take away that impression? The current text says (with additional qualifying details) that "an analysis ... showed" this was impossible. A statement supported by sources. (I believe that is your language.) He hasn't suggested an alternative text, he just wants to remove any hint of impossibility. It seems that wants to leave the impression that such travel IS possible. He is using the criterion of length to steer the content more favorably.

As to "biased selections", I have run out of time to engage in detail so I will summarize in a singe word: bullshit. As an added gloss: I am disappointed it appears you are not paying attention. ~ J. Johnson (JJ) (talk) 23:57, 25 September 2016 (UTC)

Please find here the sourced content on SES propagation, deleted among others by JJ without justification. Lies Appealing non-true statements does not help towards consensus.--IP202-77.69.68.33 (talk) 00:28, 26 September 2016 (UTC)

IP202, I don't understand why you are accusing JJ of lying. Perhaps he doesn't remember having deleted your sourced content on SES propagation. Or perhaps he believes for some reason that your content was not neutral. JerryRussell (talk) 01:40, 26 September 2016 (UTC)

JJ says: "He hasn't suggested an alternative text". Well, I have, and it was sourced. He might have forgotten it or reverted it without reading it, but the first makes it an unintentional forgivable lie and the second a big obstacle towards consensus. Non-neutral wording can be corrected or adjusted, and I personally welcome this, because of my english level.--IP202-77.69.68.33 (talk) 02:35, 26 September 2016 (UTC)

IP202, thanks for assuming good intentions. FYI, in English the word "lie" always implies intentional deception. I've tried again to insert your material. I also edited the example prediction telegram material to reflect my reading of Varotsos et al 1996a. The example turns out to be a successful prediction, if I'm reading the tables correctly. Please check my work? JerryRussell (talk) 03:53, 26 September 2016 (UTC)

In the "Electromagnetic variations", the inserted material related to SES transmission and SES distinction from noise is correct. The example prediction telegram is also correct. Since we speak of corrections, there are two issues left unresolved on VAN. In electromagnetics, the "Taken as a whole.." paragraph is claimed by VAN to be built on publications that violate fundamental principles of statistics ↑↑, as shown additionally by third scientists. These citations were also present on June 27 and now are missing. The second issue is with the second example, where Varotsos put the public at rest, saying that a destructive earthquake was not about to happen, to avoid panic reaction, and was accused for claiming success for a prediction he took back. On March 24, with new very strong new evidence in hand, Varotsos warned the government and two days later a destructive earthquake occurred. He was accused for not warning the public and therefore there was no prediction, but, newspapers wrote: "Earthquake Planning and Protection Organization publicly acknowledged receipt of the warning issued by Professor Panayiotis Varotsos for Elia Prefecture[ Pirgos region]" (this is found in page 105 of the book "EARTHQUAKE PREDICTION BY SEISMIC ELECTRIC SIGNALS: The success of the VAN method over thirty years" (SPRINGER 2013, ISBN: 978-3-642-24405-60) by Mary Lazaridou-Varotsos). The last sentence of the paragraph should be given a second look.--IP202-77.69.68.33 (talk) 18:53, 26 September 2016 (UTC)

Jerry, I think I will reiterate my comment about you carrying and chopping. You have totally missed the point about why the telegrams were cited. The "selection" of those examples had NOTHING TO DO with their being successful or not; that is entirely irrelevant. But you just couldn't wait to discuss it, you just had to run out and do some chopping. BTW: I remind you that IP202 was previously advised that insinuating that someone is lying contributes nothing to the discussion, so his clear and repeated accusations cannot be excused on the lame basis of "bad english".

IP202: As you have once again reduced a discussion to a personal attack, let's take a look at some facts about just who is "lying'. You started this discussion "VAN sections too long?" complaining that "JJ kept on building anti-VAN content without listening and by reverting all IP202 edits", that without Jerry's timely intervention "JJ's hill would be a mountain now." Which is just out and out bullshit. But more to the point: your comments are contrary to fact. In fact, through June, July, and August it was YOU (or your associate) making the article longer, with multiple large edits. (E.g.: here, here, here, here, here, here, and here, and those are only the larger edits for June.) Of your very few and insignificant deletions, the largest appears to be this removal of some embarrassing details you would rather suppress. In blaming me for the growth of the VAN section you have it backwards: it is due primarily to YOUR edits. Your statement was intentional, and it is FALSE. Which is to say: you lied.

You also state that I was "reverting all IP202 edits", and in an earlier comment, that I have "excluded, without consensus, all VAN published responces...." All? You have quite disingenously implied that I am the only editor opposing you (just as you have insinuated that all VAN criticism is due to Robert Geller). In fact, I have not reverted "all" of your edits, and even in reverting some of your disputed additions I was joined by Ismabard Kingdom, Volunteer Marek, and William Connolley. Again, your statement is false: you lied, and repeatedly.

And then you claim that my statement that you had not "suggested an alternative text" is a lie. It seems (you never can spell things out directly, can you?) you would claim that my statement is false because you linked to earlier version of the article which just happened to contain some of your "alternative" text. To which I say: bullshit. (You're so full of it.) Your opening comment here referenced size, and said nothing about alternative text. You still have not presented any alternative text (certainly not here), and to claim otherwise is to pull the same smoke and mirrors trick you do with predictions: point vaguely without any definite statement, then make it up afterwards.

Are we clear now on who is lying? ~ J. Johnson (JJ) (talk) 21:38, 26 September 2016 (UTC)

Nobody is "lying" here, we just all have different points of view. There seems to be some confusion about what "alternative text" was asked for, or provided, when and where.

As to reasons why the article is too long, isn't it the fact that both of you are correct? IP202 and other Greek editors started the ball rolling by making many additions to the text, beginning in June. Much of their added text was reverted, but some was kept, and then JJ added more text too. I've played my part as well, right? And I'm not advocating deleting any of it from Wikipedia, just asking if we can offload some of the detail to the VAN method article.

JJ, are you asking for a proposal for an alternative, shortened text for this article? JerryRussell (talk)

Jerry: you are quite the Pollyanna, in not seeing that whether anybody is actually lying is rather secondary. IP202's point of view, and his clear insinuation, is that I am lying. Clear enough that you asked why he was accusing me of lying. Remember? And have you forgotten IP202's previous history of accusing me of lying?

It is not at all a matter of "some confusion", as there is no confusion at all: he is calling me a liar. If it was a matter of confusion then the proper response is to ask questions, or offer information. But no, he just starts insinuating that I am a liar. That shows he really isn't here for any kind of civil resolution.

No, I am not asking for alternative text. As I commented just below, your recent edits on VAN, done at IP202's behest, were all additive, showing that "too long" isn't really an issue here when VAN is being promoted. If "too long" is really an issue, then just revert all of the recent additions, which only promote a VAN sensitive point of view. Note that my original text was much shorter, because I summarized a lot of material. Where I have expanded text about VAN it is been largely because IP202 complained of a lack of sources, etc. Having engendered this expansion of everything touching VAN it is dishonest dealing to complain of the result. The best "alternative" is to just go back to the original. ~ J. Johnson (JJ) (talk) 21:38, 28 September 2016 (UTC)

JJ, do you see that IP202 retracted his accusation that you were lying? He admitted that he might have been complaining about an unintentional mis-statement. And when I tried to de-construct the situation further, and identify what had been mis-stated, I found that there was extensive confusion about which statements were being debated.

Are you standing by your contention that IP202 is, in fact, a liar? If so, shall we deconstruct that and see if we can clear up the problem? Perhaps IP202 has made false statements unintentionally, and perhaps he would be willing to clarify or retract his statements.

It's not correct that all my recent edits have been additive. For example, with respect to the two example coordinates ("200 km W", etc.) I began by deleting that sentence. It's only after you objected that I then restored the example and added more remarks for clarification. Now I've folded my additional sentence back into the footnote, and I don't really care if that sentence ever comes back into the main text.

With respect to the new material on SES propagation mechanism through the earth, and EM interference, I'm sure IP202 would be just as happy if that material were deleted along with the derogatory statements about the impossibility of VAN signal transmission and reception. In general, moving material into the VAN article would need to be done in a neutral way, otherwise we would lose whatever consensus we've developed.

The material about M&S is proposed for the section on evaluation. Nobody is concerned about that section's length. JerryRussell (talk) 23:40, 28 September 2016 (UTC)

Sorry, no I do not see that IP202 has retracted his accusation of lying, for the simple reason that (so far) no such retraction exists. If he wants to retract anything, then he should say so himself, and strike-out the insinuations. What he said (02:35, 26 Sep) is that I "might have forgotten or reverted" his supposed alternative text, as if he really had proposed some alternative text (not true), and therefore it is my fault that he insinuated I was lying. So while he fake-graciously excuses me for "an unintentional forgivable lie", he still claims a "lie".

My contention that IP202 is a liar (above) is derived from the same principle he seems to use: that any utterance of a false statement is a lie. I have shown that has made false statements. Therefore, by his own standard, he is a liar. I would not have troubled, but that he has a history of calling me a liar, and does not yet seem to accept the disutility of personal attacks.

Sure, not all of your recent edits have been additive. But more so than not, and also selective. E.g., the deletion you mention relates to the criticism that VAN is so allergic to. Which gets back to my previous comment about selective application of general rules, and my comment (next section) that being "too long" doesn't seem to apply to VAN-positive material.

I point out that your characterization of the "statements about the impossibility of VAN signal transmission and reception" as "derogatory" is so very non-neutral as to raise a question of your impartiality. ~ J. Johnson (JJ) (talk) 22:55, 29 September 2016 (UTC)

JJ, I didn't mean to convey any impression of non-neutrality, and I can see how my choice of that word could be seen as non-neutral. I have stricken it out, and will discuss further below in the appropriate section.

According to Lie and other online dictionaries, and common English use, a lie is a statement that the stating party believes to be false and that is made with the intention to deceive. Therefore IP202's formulation "an unintentional forgivable lie" was a bizarre contradiction in terms, such that I felt the appropriate response was to give IP202 an English lesson. I think his meaning was clear, though, that he felt the false statement (if any) was unintentional and forgiveable. (For whatever it's worth, I do also presume IP202 would stand by the view that your position is not neutral with respect to VAN.)

Likewise, thanks for the clarification that you are only claiming that IP202 has made false statements. I'm quite sure he has made false statements from time to time. However, in the particular case you mention, I still see only confusion. With respect to "proposed alternative text", I think IP202 meant the link he provided to an earlier version of the article (which required the reader to figure out what text IP202 was referring to), whereas I take it you were looking for an easy-to-find, new and specific proposal. So I stand by my position that neither one of you has even made a false statement in this context. There's nothing to see here but a failure to communicate. JerryRussell (talk) 14:56, 30 September 2016 (UTC)

Jerry, the problem with your "derogatory" is not the saying of it per se, but that you evidently have that viewpoint. It showed an over-readiness to fall in-line with the VAN/IP202 viewpoint without giving it enough thought.

In regards of "lie", it doesn't matter what IP202 thought; what he did is accuse me of lying. That is uncivil, and more particularly, it is personal attack. For you to try to mollify matters does not restore civility; it only encourages him to continue with his incivility, as "bad english" is now his "get out of jail free card".

Note that I am not saying "only" that "IP202 has made false statements." I am saying that by his own standard he is a liar.

As to "a failure to communicate": yup. In a large part it comes from people vaguely waving their hands around without actually pointing to what they are talking/complaining about. I am not a mind-reader (are you?), and anyone that can't make a clear statement of what they think they are thinking has no basis to complain of a communications failure.

That my position "is not neutral with respect to VAN": well, of course. That is the key issue in ALL of this: IP202 wants everything to be "neutral" with respect to VAN, with an assumption of equal validity. But that would be a WP:FALSEBALANCE, and not the standard we use. The standard we use is relative to mainstream views. Except that he doesn't want even the false balance of equal validity and equal presentation, he wants the VAN pov boosted (for balance!) while the criticism is removed (because the section is too long!). (E.g., see his latest comment.) ~ J. Johnson (JJ) (talk) 22:04, 30 September 2016 (UTC)

Regarding the civility question, my personal view is that long-standing experienced editors should be held to a higher standard of civility, while non-english speakers and occasional users (even SPA's) should indeed be cut some slack. Although IP202's statements are not as clear as they should be, I generally don't feel I have any problem understanding what he's saying. My interpretations could be wrong from time to time.

As we've mentioned earlier, JJ, I believe if you want to take IP202 to AN/I, you would probably get a block; it's entirely up to you. But as I'm sure you're aware, your conduct would also be subject to scrutiny.

About neutrality and WP:FALSEBALANCE, I will start a new section below. JerryRussell (talk) 01:10, 1 October 2016 (UTC)

Why are you making excuses for him? Despite several admonitions he has continued to make personal attacks. "Non-english" does not excuse this, especially after it is explained to him. And you might note that on several occasions he has implied that he is an experienced editor. ~ J. Johnson (JJ) (talk) 23:50, 1 October 2016 (UTC)

Inasmuch as our time is valuable, and I've already stated why I am making excuses for IP202, you may review the page to find my answer. Or perhaps you might be able to remember. Or perhaps you'll go right on thinking whatever you want to believe about that, regardless of what I might say. JerryRussell (talk) 00:51, 2 October 2016 (UTC)

I seem to recall you saying he had made a valuable contribution, but I don't recall that you ever pointed out what that contribution might be. Even if he had made some minor contribution, how how does that justify his persistent incivility? How about making a clear, definite statement so that I and any passers-by can know, without having to go through all the malarky? ~ J. Johnson (JJ) (talk) 05:08, 2 October 2016 (UTC)

JJ, on several occasions IP202 has posted links to information online, or even substantial excerpts from articles hidden behind paywalls. I don't have a big budget to buy access, nor convenient access to a library, so this has been very helpful to me in understanding the issues. Also, he has frequently pointed out areas where (in my opinion) the treatment of VAN has not met NPOV criteria.

I have not attempted to "justify" IP202's incivility on the basis of his contribution. I have attempted to explain it ("make excuses") on the basis of his limited English skills, and possibly different cultural and legal standards in Greece. JerryRussell (talk) 16:31, 2 October 2016 (UTC)

Since you've reminded me about IP202's indication that he has been a long-term Wikipedian, I went to look up that thread. Here's the exchange:

IP202: I have not seen that nerve before anywhere in Wikipedia, all my years here.--77.69.80.202 (talk) 18:35, 27 June 2016 (UTC)

jj: All your years? Your earliest edit as "77.69.80.202" is 12:25, 5 June 2016, not even a month ago. If you have edited from other addresses, or have other history here, you should disclose that to avoid any taint of sock puppetry. ~ J. Johnson (JJ) (talk) 22:09, 27 June 2016 (UTC)

At the time, I remember parsing this in lawyer-like fashion, and concluding that perhaps IP202 had been using Wiki for the purpose of reading articles, and had never seen such bias. Or that maybe he'd just been reading talk pages. But, that does seem like a stretch to give him the benefit of the doubt. It does read more like a confession that IP202 has edited under multiple IP accounts, or under an earlier registered account.

WP:VALIDALT gives a number of legitimate reasons why an IP editor might use multiple accounts for editing, while WP:ILLEGIT lists inappropriate uses. Since IP202 has declined to declare his earlier edit history, this would be another possible topic for an AN/I investigation.

I'm disappointed to see that IP202 has disregarded my recommendation that he defer from editing the article. So much for any gentleman's agreement that existed in my imagination. JerryRussell (talk) 17:59, 2 October 2016 (UTC)

Is earthquake prediction intrinsically impossible?

Even aside from SPA and advocacy issues, IP202's English is really bad enough that he shouldn't be editing. I've reverted him, and attempted to clean up a grammatical error that existed in the paragraph. Basically I think the paragraph in question was correct and neutral as it stood. JerryRussell (talk) 18:19, 2 October 2016 (UTC)

I do not wish to use an account for editing the article. IP202 is a single-purpose "account" and there may be an essay on this, but I have not argued talking to "another myself" in this page to affect consensus. Regarding my bold edit, basically, I think the paragraph in question was proven as non-neutral as it stood.--IP202-77.69.68.33 (talk) 09:48, 4 October 2016 (UTC)

IP202, I want to drill down on the edits to this paragraph. The first sentence originally read Earthquake prediction may be intrinsically impossible. You changed that to read Earthquake prediction is claimed to be intrinsically impossible. So the original was in Wiki voice, which would be appropriate if this is the consensus among seismologists. Based on everything I've read, this is probably true: everybody is at least concerned that it might be impossible, especially if by "prediction" is meant "reliable and skillful prediction" for purposes of evacuating cities. Whereas you changed it to speak in source voice, as if this was a controversial proposition. Whether intentional or not, your change to the first paragraph was strongly in favor of VAN, and IMO rendered it in violation of WP:FALSEBALANCE.

The last sentence read "These theories and their implication that earthquake prediction is intrinsically impossible are still disputed." You changed it to "These theories and their implication that earthquake prediction is intrinsically impossible are disputed." If there's any difference in meaning between those two formulations, it's certainly subtle. But, I thought that both formulations went beyond the citation from 1999. I have no idea whether there's an ongoing dispute about this topic. Accordingly, I put that sentence in past tense, These theories and their implication that earthquake prediction is intrinsically impossible have been disputed. If anything, I think that change is also unfavorable to VAN. That is, the implication is that no one has argued specifically against the intrinsic impossibility of EQ prediction since 1999. And one can hardly claim that VAN successes are a counter-argument, inasmuch as VAN has not demonstrated that they have overcome their problem with high false alarm rate. JerryRussell (talk) 17:16, 4 October 2016 (UTC)

This is not a paragraph on VAN and I cannot follow the focus on VAN above. The big deal was with the "still" word, meaning that any time now or in the near future the possibility point of view will not exist, for any method or combination of methods. I am happy as it is now. I like the poetry of "may" or "however", I agree not to be too strict on the encyclopedic tone.--IP202-77.69.68.33 (talk) 21:08, 4 October 2016 (UTC)

IP202, I agree that the paragraph applies to any prediction method. I was just using VAN as an example, and wasn't meaning to imply that the interpretation needs to center on VAN.

It's interesting how you interpreted the word "still". I think of it to the contrary, as emphasizing the eternity of the process. As in: "Even still to this day, Christians celebrate Easter." Anyhow I'm glad you're happy, I hope JJ will be happy too. JerryRussell (talk) 22:24, 4 October 2016 (UTC)

Ah, but "intrinsinc impossibility", even if not explicitly about VAN, certainly touches them. Because if EP is impossible, then their efforts are meaingless. This is why IP202 has been messing in that section, to ensure that all readers get the idea that "intrinsinc impossibility" is disputed. Recall his previous edits, where he wanted to expressly exclude "negative views" on this (see previous discussion at #"Classification" and a non-NPOV). ~ J. Johnson (JJ) (talk) 00:52, 5 October 2016 (UTC)

Two more unresolved issues?

IP202, thanks for checking my edits. Again I'm hoping that JJ will join our consensus. With respect to "two more unresolved issues", here are my views.

In the paragraph following "taken as a whole", the article states that VAN had an opportunity to reply to their critics in the GRL and Lighthill publications, and citations are given for both of those volumes. This seems sufficient to me, and it might be undue weight or citation overkill WP:OVERSITE to provide more detail.

In the paragraph about the Pirgos 1993, the main point of the paragraph is that Uyeda claimed that lives were saved, but Chouliaris & Stavrakakis disputed this. It's already mentioned that Varotsos' public statement of Feb 26 was unclear as to whether any predictions were being withdrawn, and that further predictions followed after March 5, which were not released to the public. So again it seems to me that the existing paragraph is complete & accurate, and further detail or more citations are unnecessary. However, the information that Varotsos issued a warning to the government on March 24 might be relevant -- but only if there's any evidence that the government took some action based on that warning, and that lives were saved by government actions. Does L-V 2013 mention anything to that effect? JerryRussell (talk) 20:05, 26 September 2016 (UTC)

It was more like a feeling that these topics were not really ever discussed. I agree there is balance as a whole, after your important additions on SES. I will study Lazaridou tomorrow to find out about possible government actions.--IP202-77.69.68.33 (talk) 21:55, 26 September 2016 (UTC)

The "important additions on SES" showing that being too long doesn't apply to VAN-positive material, further demonstrating the continuing promotion of VAN. ~ J. Johnson (JJ) (talk) 22:04, 26 September 2016 (UTC)

Example VAN predictions

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

JJ, thanks for finding the link to that old discussion. IP202's comment was: The selection of the two examples has been done with unjustified criteria, using only secondary sourced material. These examples are equal to examples of success that are omitted. I replied, in full:

The selection of two examples "such as (typically) a prediction of an earthquake at either 300 km "N.W" of Athens, or 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit" is intended, as it says, to be a typical representation of problems with VAN predictions, namely a lack of precision. The selection could be criticized as "Original Research" but then again it might be equivalent to an obvious application of simple arithmetic. I would have no objection if you were to pull out a couple of examples of 'successful' predictions in which the location, time and amplitude of the quake were all predicted accurately in advance. On the other hand, I would also have no objection if someone were to strike the text about the two predictions as WP:OR, but it's better if we can agree with JJ here in the talk section so that we don't get into edit warring.

There was no reply to that from you or IP202 in the paragraphs that followed, so I asked again:

Also, the selection of two examples from a huge list seems to be a possible violation of WP:OR. I suggested that maybe IP202 could select a couple of counter-examples, but that might be just more OR, and also lengthen an article section that's already too long. How do you think this should be resolved?

Neither you, nor IP202, nor anybody else, ever commented about this again -- until IP202 remembered it in the above thread.

At Wiki, there's always room for editorial judgment about the correct application of policies. I don't question anyone's good faith about their opinions regarding whether the two examples are prejudicial against VAN, or whether the selection was OR, or whether the examples might actually be helpful to the reader. I have an opinion and I've stated it and acted on it, but I can only look to the consensus process to resolve any dispute.

JJ, if this is really important to you, why don't we ping a few more editors and get their opinion? JerryRussell (talk) 00:25, 23 September 2016 (UTC)

These remarks you have quoted are from the #Libel on VAN 1983-1995, opened by IP202 (15 Aug.) with four bulleted paragraphs which (lacking any kind of explanatory comments) we can only surmise were intended to support his previous complaint that the article (or I personally) libeled or defamed VAN. The second paragraph states: "The selection of the two examples has been done with unjustified criteria, using only secondary sourced material. These examples are equal to examples of success that are omitted." The identification of these "two examples" with the cited telegrams comes from your subsequent comment. You opined that the selection "could be criticized" as OR, or (logical alternative!) "might be equivalent to an obvious application of simple arithmetic". And while you had "no objection" one way or the other, you also allowed that "it's better if we can agree with JJ here...."

And then after another thousand words about length, liars, and insinuations, we dived down a five-thousand word rabbit hole on #Are the VAN telegrams a complete record of attempted predictions? Please excuse me if I overlooked that we had not sorted out whether IP202 was complaining about unspecified "unjust criteria", or an unrepresentative result, or something else. Especially as IP202 has not provided any elaboration or explanation of what he meant.

So if you were under any misapprehension that there was consensus for deleting those examples: sorry, no, you were wrong. And now that you know, and especially as your allegation (not IP202's!) of original research has been fully deconstructed in the section above (#IP202 apparent COI?), perhaps you would see to correcting your misstep. ~ J. Johnson (JJ) (talk) 23:51, 23 September 2016 (UTC)

I don't feel that the allegation of original research has been "fully deconstructed" yet. I haven't been able to read one of the two sources, and I haven't studied the other one carefully enough to form an opinion about whether the two examples are representative or not. And as I mentioned above, I also feel that the article section is too long, and that this is "undue weight". I'm sorry that I was mistaken about a consensus. If you revert me, I won't edit war about it. JerryRussell (talk) 00:24, 24 September 2016 (UTC)JerryRussell (talk) 02:56, 24 September 2016 (UTC)

As you haven't studied the point well enough to have an opinion on it, why did you presume to act? I remind you that I do have those articles, and I have studied them (and others). Your act of deleting those examples says, in effect, that you think your unstudied opinion is not just better than my studied (and sourced) opinion, but so much better that there was no point in even bothering to ask (me, or anyone else) about this. Which is pretty insulting, so perhaps you understand why I am a bit irked about this.

As to what is being deconstructed here, you need to make up your mind. Is it:

1) IP202's complaint that there was "selection ... with unjustified criteria, using only secondary sourced material"?

2) your variation, that the examples are possibly "unrepresentative?

3) your opinion that the examples could be criticized as "Original Research"?

Regarding the last, oh yes, it could be criticized as OR, but that is not quite the same as showing that it is OR. Note that the essence of original research (from the WP:OR nutshell) is this: "all material in Wikipedia must be attributable to a reliable, published source." As the examples are (were) attributed to "reliable" (ostensibly), published sources, there is no question of OR. That criticism is entirely unfounded. Note that the requirement is not "... a published source available to Jerry Russell", just "... a published source." Period. What else needs to be said?

IP202's complaint can be dismissed out of hand because he makes no showing that any "unjustified criteria" were used. (Indeed, how does he even know what criteria were used?) He asserts that "only secondary sourced material" was used, which is clearly a factual error in that the material is clearly attributed to the man himself. I don't know how much more "primary" one can get.

As to the rest I direct you right back to the preceding discussion just above at #IP202 apparent COI?. I will note (regarding your example) that if (say) an article had well-sourced statements that babies were being nailed to crosses at "some temple fo Kali" (gruesome yes, but hardly more then your "images of Kali eating babies"), and there were a hundred images of such crosses, "selecting" two of those images is not per se misrepresentation. (And most certainly not OR, no matter the selection criteria used.) Your concern that I have in someway "chose[n] two images showing Kali eating babies" out of hundreds showing Kali doing good deeds presumes that the two images (telegrams) were in fact non-representative. But, in fact, they are not. In all of their publications VAN use this indefinite "N.W. of Athens" kind of "location". Note also that IP202 has totally failed to provide any counter-examples, as you had suggested.

In summary (and including the preceding discussion) I say these objections to the telegrams are bogus. Enough? ~ J. Johnson (JJ) (talk) 22:29, 24 September 2016 (UTC)

Being spurred to greater effort, I went to read more carefully at Varotsos et all 1996a in Lighthill. Summary predictions for ~1987-1995 are in Table 1, pp. 50-51 and text of predictions ~1993-95 in Table 2. Only about 1 in 4 of the predictions is given in the form of an "OR" of two different possible EQ events, and the vast majority give full coordinates and specific magnitude estimates. The footnote says that the example prediction is from "Telegram 39, issued 1 September 1988, in Varotsos & Lazaridou 1991, Fig. 21, p. 337" but this same prediction appears in Table 1 as telegram 10. According to Table 1, there doesn't seem to be any ambiguity that two different EQ are being predicted, with two specific magnitudes and locations, and no confusion about which location goes with which magnitude estimate. So it's looking to me like IP202 is correct, that there is some selection bias in this example. JerryRussell (talk) 01:19, 25 September 2016 (UTC)

I'll review that when I have more time. But off the top it does look like you're still wrapped around the axle. ~ J. Johnson (JJ) (talk) 21:43, 26 September 2016 (UTC)

JJ, above you wrote: Jerry... You have totally missed the point about why the telegrams were cited. The "selection" of those examples had NOTHING TO DO with their being successful or not; that is entirely irrelevant. But the selected examples don't actually succeed in demonstrating the alleged spatial vagueness of the VAN predictions. With respect to your point above In all of their publications VAN use this indefinite "N.W. of Athens" kind of "location": in fact there is nothing at all indefinite about a location "300 km NW of Athens." That formulation specifies a particular point on the map, which can also be defined in degrees of latitude and longitude. In Varotsos et al 1996a, all the coordinates for all the predictions are given in degrees latitude and longitude. Some of the telegrams give two different location predictions, but it doesn't seem that they mean the EQ could be anywhere in the range between the two locations given. My original complaint was that the majority of prediction telegrams from VAN were not in this "either/or" format, most of them specify just one location, so the use of this example in "or" format seems to be cherry-picking. But neglecting that issue, I feel it's clear that most VAN predictions cannot be criticized for a lack of spatial precision; and thus the example you gave doesn't really work for that purpose. I believe it's correct, though, that VAN doesn't put time limits on their predictions. Also, if they are just specifying a single point on the map, it's open to interpretation whether any particular EQ is "close enough" to be considered a hit. JerryRussell (talk) 18:26, 27 September 2016 (UTC)

You are so wrapped around the axle it is hard to decide just where to start cutting you loose. You opened this section with IP202's comment (axe) that "The selection of the two examples has been done with unjustified criteria, using only secondary sourced material". (An assertion he made in #Libel on VAN 1983-1995, where its relevance was never established.) But while you assert "unjustified criteria", you don't actually specify what it is. (Or what sources were used.) You're just waving your hands around.

You have also asserted that "the selection of two examples from a huge list seems to be a possible violation of WP:OR." Which is just more bullshit. In the first place, there is (ha ha, fooled you) no "huge list". The examples were "selected" from a set of exactly two. From which I "cherry-picked"... the entire list.

Confused? Of course, because you haven't been paying attention. So please: PAY ATTENTION. The examples are just that: examples, and particularly of VAN's use of vectors to specify location. In the source used (VL 1991; I didn't use Dologlou 1993 because at that time I didn't have it) there are only two facsimile images of prediction telegrams. Now we don't know what criteria the authors used to select those two telegrams of however many they may have, but I think we can be confident of at least two things: 1) Geller didn't select them, and 2) VAN wouldn't select disfavorably.

You seem to think it is significant Varotsos et al. 1996a (CRV) used lattitude/longitude to indicate location. It seems you have failed to notice that their Table 1 is a recompilation of other sources, including VL91 and Dologlou 1993. Which all use vectors, not lat/lon. The significance of Table 1 is that VAN changed format. And (if I recollect correctly) without any mention of that change, or why. Which is why the original telegrams need to be checked, instead of relying on someone's say-so as to what they say.

You also say that 'there is nothing at all indefinite about a location "300 km NW of Athens."' Sorry, you are wrong. (Which you might have noticed if you had read any of the sources that criticize VAN on the point of indefinite location.) The matter is very much (as you allow) 'whether any particular EQ is "close enough" to be considered a hit.' The problem is that VAN are never definite as to what is "close enough", leaving it to the critics to infer what they might mean from their other publications. Which VAN then quibble about.

And that is about all I have time for today. ~ J. Johnson (JJ) (talk) 21:51, 28 September 2016 (UTC)

JJ, thank you for clarifying that the 9/1/88 telegram or telegrams were singled out for reproduction in VL 1991, and thus that your selection was not "original research". You could've just said so, instead of going into this long conversation about a hundred pictures of Kali temples. I explained a long time ago that I don't have a copy of VL 1991 and have been relying on the presentation in Varotsos et al 1996a, which is also cited in your footnote. The translation between vectors and lat/lon seems to be a simple equation. I don't know why VAN's use of vector coordinates in one document, and translating to equivalent rectangular coordinates in a second publication, would be something they would need to "explain".

If the problem is that VAN aren't definite about what is "close enough", then maybe the article should just say so. JerryRussell (talk) 22:47, 28 September 2016 (UTC)

Jeesus, Jerry, the "hundred pictures of Kali temples" was your metaphor. And the conversation could be greatly shortened if you could just specify at the outset exactly what your point or complaint is. As it is, even now you raise about three different points. But I'll try to cover them all.

I know you don't have VL 1991, but you really should accept that you have no grounds for arguing about material without going back to the source. Where you might have a question, or even a general concern, I will try to assist, but you need to focus on what that concern is.

The significance of vector/lat-lon conversion is not that it can't be done, but that is was done without note. It amounts to an alteration of the original data, as it allows the precision of the data to be "improved".

I believe that at some point the article did cover, however inadequately, and seemingly an eon ago in edit history, the indefinteness of "close enough". But perhaps this got lost in the tension between general summarization and detailed specification. I should like to rewrite that entire section, but all my time gets soaked up here. ~ J. Johnson (JJ) (talk) 23:43, 29 September 2016 (UTC)

Actually you suggested "a hundred pictures of temples" as a metaphor, and then I added the part about Kali and the babies, as an illustration of the fact that choosing two images from a hundred can give rise to editing controversy. I was working from the basis that IP202 had complained about the selection of the example, and I didn't see any basis within Varotsos et al 1996a for selecting that example out of a long list.

So it turns out that VL 1991 singled out those examples. None of us know for sure why they chose them, but a reasonable conjecture would be that in fact, they nailed that particular prediction very precisely. This is all rather ironic, as IP202's original complaint was that a successful prediction could have been chosen for illustration.

I can't be sure because I don't have VL 1991, but based on the information in Varotsos et al 1996a, it appears that your text formulation inferring that VAN was giving a range of locations was incorrect, and that the two predicted locations were in "either / or" format. So I rephrased the presentation based on making that clarification. Do you still have any objection to the text as it currently appears in the article? I put the information that the prediction was a hit, into a footnote; which I think actually is an improvement, as it focusses attention on the intent of the example in the context.

Is there any secondary source that argues that VAN did, in fact, "improve" the precision of the data in translating from vector to lat/lon? In the one example I checked, the conversion seemed to be done correctly, with no clear implication of greater precision. As to the indefiniteness of "close enough", you might very well be correct that an earlier version of the article made this point. In the existing version, it should be easy enough to clear this up, without resorting to a complete re-write. JerryRussell (talk) 15:19, 30 September 2016 (UTC)

Yes, you were working from the basis that IP202 had complained about, except that IP202 never gave a clear, distint basis for what he was complaining about. Way back then he started complaining of "libel", then (in regard of this particular point) he said "unjustified criteria", without ever explaining, pointing, or even hinting as to what those criteria might be. Then he complained that the two examples were "added only to support Geller's point of view", after which YOU back-filled with "Original Research". Only now are you clarifying IP202's meaning (you are a mind-reader!) as "a successful prediction could have been chosen for illustration." Only you still don't get it: whether that prediction was successful, or not, is ENTIRELY IRRELEVANT (!!!) to the purpose of illustrating a typical telegram. In fact, I could NOT have chosen another, as, in using both of the telegram images available in that source, THERE WERE NO OTHER EXAMPLES.

Why is that so difficult for you to grasp? Are you paying attention??

At least one source discusses the implications of using vectors. It would be salubrious to find it your self.

It is quite difficult to compare relative indefinitudes. But in your checking consider this: what is the difference between a point 300 km NW of Athens and a point 300 km N of Athens. Take half of that as the radius that defines a prediction area about the point. But wait, Athens is a big place, so do they any where specify an exact origin for the vector? If not, add half the width of Athens (as an approximation) to the previous radius. Now calculate the area included in a tenth of a degree of latitude and longitude (the precision implicit in Table 1). If you calculate it as a rectangle be sure to calculate the vector-derived area as a rectangle. Question: what is the proportion of one area to the other? ~ J. Johnson (JJ) (talk) 23:18, 30 September 2016 (UTC)

Hello JJ, what IP202 said was "These examples are equal to examples of success that are omitted. By 'success' I construe that he meant predictive success. Regarding your in that source, THERE WERE NO OTHER EXAMPLES.: How many times have I told you I don't have access to that source? I don't question your clear assertion that there were no other examples, but on the other hand I can't independently vouch for it. What I did was to look at Varotsos et al 1996a, where there were many other examples.

I agree that relative indefinitudes are difficult to compare. I don't agree that taking half of the radius between two predictions should define the intended precision of the vector predictions, nor that it's appropriate to assume that the implicit precision of table 1 was the intended precision. If you know of a source that discusses these relative indefinitudes, please let us know, and perhaps post a short excerpt for discussion. My time is valuable too, I'm not going to go on a search for information that you already have at your fingertips.

I don't agree that the selected example really says anything at all about the comparative relative indefinitude of VAN predictions. So it's equally as irrelevant as the success or failure of that particular predictive example. That's why I initially tried to cut the whole thing. Now the information about the predictive success is in a footnote, and it can stay there as far as I'm concerned. JerryRussell (talk) 01:28, 1 October 2016 (UTC)

In Lazaridou 2013 book, in the table of contents one can find the cases VAN consider most significant throughout 30 years of VAN method apply. Pirgos 1993 is indeed claimed as a successful prediction.--IP202-77.69.68.33 (talk) 04:32, 1 October 2016 (UTC)

Jerry, my time is also valuable, and also limited, so I am getting just a little exasperated at your continuous disputation of these points. I don't mind getting down on a few selected points to test my understanding and competence, but it gets tiresome when you start disputing everything. (Like where you "don't agree" regarding the "half radius". I could explain it, but then you would stick at something else. Right?) Perhaps you would consider listing a few points to examine, instead of always running of to another point.

I earnestly urge you to be more deliberate, and more attentive. E.g., you claim that Vartosos et al. 1996a had "many other examples." Oh? Can you give me the page numbers where they have facsimile images of prediction telegrams? My recollection may be faulty, but I don't recall seeing any. I suspect you may have been thinking of "examples" of predictions, such as in their Table 1. At any rate, where I said there were no other sources why is it that you seem to have not picked up where I specified (referring to VL 1991)"telegram images available in that source"? Skipping that is sloppy, and only gets us more entangled. ~ J. Johnson (JJ) (talk) 23:46, 1 October 2016 (UTC)

Of course, JJ, I was thinking of "examples of predictions." I don't understand why it's important that the predictions be presented as actual facsimiles of telegrams, or why you put that specification in the footnote. For me, that has never been what this discussion has been about.

Do you have any complaints or concerns about the way the material is now presented? Are there any errors in my edit to that material? JerryRussell (talk) 00:55, 2 October 2016 (UTC)

Well, that explains it all, I was talking about illustrations of telegrams. (Didn't I say "telegrams" a couple of times?) So now you see that the selection of telegrams for purposes of illlustrating the form and wording of said telegrams was 'not faulty? And we can close this complaint of "unjust criteria" as unfounded?

Facsimile images (need I be more precise?) of the original telegrams are needed because they are the only a priori record of any prediction, and without seeing the originals, without alteration or interpretation, it is impossible to verify what was predicted. Like whether they used vectors or lat/lon for location, or specified the magnitude scale. Which does make a difference in evaluating the predictions.

I have a number of concerns, particularly with your understanding of the material. I'll tell you about them later, when I have more time. ~ J. Johnson (JJ) (talk) 05:52, 2 October 2016 (UTC)

JJ, before this whole kerfuffle started, here's what the text of the article said: Much of the controversy with VAN arises from this failure to adequately specify these parameters, such as (typically) a prediction of an earthquake at either 300 km "N.W" of Athens, or 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit. So can you perhaps understand why I thought we were talking about predictions, rather than their particular instantiation in telegrams? If you were trying to show that VAN was dishonest in translating from vector coordinates (in the telegrams) to rectangular coordinates (in tabular representations), neither the article text nor the footnote spelled this out in any clear fashion. And if that's your point, you need to find a secondary source that makes that point.

I didn't ask if you have concerns about my understanding of the material. It should be obvious to anyone that I'm a lay person regarding this topic, and that the amount of reading I've done is minuscule compared to the amount of material that has been published. What I asked is whether you have any concerns about the text of this sentence in the article as it's currently written? The text now says: Some of their telegrams include predictions of two distinct earthquake events, such as (typically) one earthquake predicted at 300 km "N.W" of Athens, and another at 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit. There is an additional footnote indicating that one of these two predictions was indeed materialized as an EQ in the predicted location. JerryRussell (talk) 16:44, 2 October 2016 (UTC)

All the more reason you should query before boldly leaping forth.

Jerry, the text in the telegrams IS "the particular instantiation" of the predictions. Where do you think the data in the tables come from?

Please note that this "kerfuffle" actually started when IP202 said (5 Aug) the examples were selected with "unjustified criteria", but without any elaboration as to how they are unjustified, or even what criteria he thought were used. At which point you started imaginatively filling in his blanks, and we spent a lot of time chasing the "selection" bunny. But perhaps now you understand that the subject of the sentence in question is VAN's failure to adequately specify the prediction parameters? That the sentence illustrates these problems in regard of location, magnitude, and time with specific wording from from the source, and that ALL of the published telegrams use the same wording?

As to the current text: it is correct, but misses the point. That VAN often doubled their target area with these alternate locations is secondary. The primary point is that the granularity of their vector direction is pretty coarse, as (with one known exception) they use only the cardinal and intercardinal directions. Their vector covers a large arc of azimuth, which gets bigger as the distance increases. Mulargia & Gasperini [1992, p. 33], allowing them 16 directions, calculated that at a distance of 300 km the azimuth has a precision of about +60 km.

To get back to any earlier subdiscussion, the problem with VAN's conversion to lat/lon (in Table 1) is that the tenth of a degree is implicitly the precision (standard scientific and engineering practice), which would be on the order of +5 km. That is a big difference, and quite aside from any question of dishonesty (which I was NOT trying to show) it makes a big difference in evaluating the results.

Aside from any issues of wording, are we done with these issues of "unjust" selection, OR, etc. regarding these examples? ~ J. Johnson (JJ) (talk) 00:07, 5 October 2016 (UTC)

Mulargia & Gasparini's precision estimate of 60km is more or less consistent with the idea of a 120 km radius. I think I remember somewhere seeing that VAN took credit for an EQ centered 120 km from their prediction. I gather the predicted locations are based on "sensitivity maps" for each of the VAN monitoring stations. Those computations presumably yield very precise locations for the predictions, which are then (optimistically, hopefully, vaguely) related to the actual locations of the EQ's. So it seems reasonable to me that VAN expresses their predictions with high implicit precision, without actually expecting that the EQ locations will be exactly corresponding.

Back on Sept. 28, I wrote: JJ, thank you for clarifying that the 9/1/88 telegram or telegrams were singled out for reproduction in VL 1991, and thus that your selection was not "original research". I'm sorry if that wasn't clear; I've been done with the issue of "unjust" selection, OR, etc. since then. JerryRussell (talk) 00:31, 5 October 2016 (UTC)

When you say the current text "misses the point", perhaps the problem is that I didn't copy the entire current text into our discussion? The article now says: A crucial issue is the large and often indeterminate parameters of the predictions, such that some critics say these are not predictions, and should not be recognized as such. Much of the controversy with VAN arises from this failure to adequately specify these parameters. Some of their telegrams include predictions of two distinct earthquake events, such as (typically) one earthquake predicted at 300 km "N.W" of Athens, and another at 240 km "W", "with magnitutes [sic] 5,3 and 5,8", with no time limit. How does that fail to make your point? I agree the earlier text made your point more obvious, but at the cost that the wording didn't correctly represent the source. (Unless the original telegrams could be interpreted as a single prediction somewhere in the range between the two vectors given. If that's the case, then Table 1 of Varotsos et al 1996a is very misleading as to the actual contents of the telegrams.) JerryRussell (talk) 00:43, 5 October 2016 (UTC)

So we concur that (aside from the phrasing of the text) the telegrams cited as examples present no problems in regard of selection, original research, or "unjust criteria" – right?

Jerry, it has been obvious from the start that you are not a scientist, and now I think we can eliminate engineering. Or it would NOT seem reasonable to you "that VAN expresses their predictions with high implicit precision ...." So allow me to inform you: precision is limited by the nature of the inputs and how they are are processed, and (generally) you cannot make more precision than you start with. It is not proper to express outputs with precision greater than inherent in the data. (Refer to any introductory engineering text, or any text that covers numerical precision.) Your presumption their computations "yield very precise locations" is entirely contrary to fact, where (another criticism; see text for sources) their locations are so broad as "cannot be practically utilized". Cranking a calculation to churn out more decimal places does NOT increase actual precision.

The point that the current text misses is the imprecision of specifying the vector azimuth with only the cardinal and intercardinal directions (N, NE, E, SE, S, SW, W, NW). The point is NOT that some predictions had alternate locations, but the location specifications (of ALL predictions) were imprecise. Nor, where alternate locations are given, is there any suggestion that the conjunction "or" means any place between the alternate locations. That comes entirely from your head. ~ J. Johnson (JJ) (talk) 21:10, 5 October 2016 (UTC)

JJ, I think you're confusing the precision of the measurements and methods, with the precision of the mapping between the prediction and the actual earthquake. VAN certainly knows the location of their measuring stations with great accuracy. I'm not certain how they compute their sensitivity maps, but I see no reason to doubt that those inputs are also highly precise. And their measurements of SES amplitudes are probably averaged over intervals sufficient to get at least a few digits of precision. So it seems plausible to me that their measurements and computations yield a highly precise prediction. And there's nothing wrong in science or engineering with expressing such a result as precisely as it's measured. There's also nothing wrong with rounding such a result to correspond to its useful social value, as a prediction of an EQ epicenter. JerryRussell (talk) 23:40, 5 October 2016 (UTC)

Of course "there's nothing wrong in science or engineering with expressing such a result as precisely as it's measured." Indeed, quite the opposite: the standard expectation is "as precisely as measured" – but no more than measured! Likewise, there is also "nothing wrong with rounding", but please note: rounding reduces precision. However, to claim precision of a tenth of a degree where you really have (at best) only half a degree is not rounding, and it is not acceptable practice. (To do so knowingly and deliberately would be dishonest, but such a result could also attributed to bad practice, indifference, etc.)

Your presumption that VAN's inputs, measurements, and computations are "highly precise" is quite dubious, and (lacking any showing of sources) quite unfounded. So where do you get that notion? You think "their measurements of SES amplitudes are probably averaged over intervals sufficient to get at least a few digits of precision", but where do you get the idea their predictions are based on averages of an ensemble of SES measurements? In the examples they have published they identify a part of a squiggely line as an SES (or whatever), and on the basis of that single instance base a prediction. Their calculation of direction (if I recall correctly) is based on the ratio of the strength of signals on perpendicular dipoles. Which appears to be just eye-balled. NOT "highly precise".

As to their "sensitivity" ("selectivity") maps, check V&L91 and the subsequent GRL article in 1996 that describes all this. They talk a lot about what constitutes "selectivity", but the actual mapping seems to be informed by what earthquakes they miss. Note they also have regional and local scale sensitivities, all of which constitute more variables by which they can finesse any misses. You should have a LOT of doubt about their precision, as they don't have much of it. Why do you think otherwise? Zen ideals?? ~ J. Johnson (JJ) (talk) 21:05, 6 October 2016 (UTC)

JJ, now you're the one who gives the impression of having no experience doing science or engineering. A "squiggly line" consists of a series of digital measurements, each of which is typically precise to a couple of digits, and the series can then be averaged, or processed using fourier techniques. Earthquakes that actually occur have epicenters whose locations are precisely triangulated. The strength of signals on perpendicular dipoles can be combined to produce a directional estimate. And in engineering, nobody gets bent out of shape if a computation is carried to an extra decimal point, beyond what the data can support. JerryRussell (talk) 15:35, 7 October 2016 (UTC)

So you think VAN's data are "highly precise" because you also think all their measurements (at least through the 1990s) were obtained digitally and analyzed (?) with Fourier techniques? Again, where do you get that notion?

Of course the "strength of signals on perpendicular dipoles can be combined to produce a directional estimate" – I said that yesterday, and it is not an issue. The issue (what you seem to have missed) is the precision that might be expected from whatever procedures they use. Which is one of the specific points of criticism: they don't specify that in their predictions.~ J. Johnson (JJ) (talk) 21:57, 7 October 2016 (UTC)

Uyeda's chapter in Lighthill (starting p. 3) claims that the SES signals were digitized and transmitted over phone lines. It's rather vague about the signal processing used, but there's a basic notion that SES is ULF, which implies some sort of spectral filtering to isolate from other frequency ranges. Then there was some sort of correlation process across observations from various dipoles. The SES amplitudes were then plugged into an equation yielding the expected magnitude of the earthquake. I have no idea whether these computations were done with desktop computers (IBM PC introduced 1981), one of the mainframes of the day, or graduate students with slide rules.

So where do you get the notion that everything was just "eyeballed", JJ? Uyeda's paper gives a very different impression. JerryRussell (talk) 03:37, 8 October 2016 (UTC)

One of VAN's earliest papers (Varotsos & Alexpoulos, 1984a) says that 18 stations "are telemetrically connected to Athens (through telephone lines) and the results are depicted on multipen recorders in the central station placed at Glyfada (GLY) about 15 km from Athens." They go on to explain how "[t]he potential difference V is measured after amplifying and filtering out frequencies higher than 0.3 Hz and the result is displayed on a strip chart recorder...." They provide a dozen illustrations of the strip charts, showing the annotations by which they manually measure their data. No mention of computers, or any kind of digital signal processing (which, if I recall correctly, was more than the computers of that decade could handle). THAT is where I get the notion the notion they are "eyeballing" their data. And your notions of "highly precise" are just quaint.

Of course Uyeda's Lighthill paper (which I don't have at hand) gives a different impression: he's trying to burnish the method. And perhaps by 1996 they had upgraded their equipment (I seem to recall mention somewhere of a PC) with – modems! But I don't recall that they have ever mentioned any significant upgrades to their equipment or processing procedures through the 1990s.

In VA 1984a they provide the underlying formulas they use, but not the algorithm by which these are applied. In VA 1984b they provide illustrations (figures 12 and 13) of how they make a "graphical determination of the epicenter" using apollonian circles – i.e., using a compass – based on the measured signal strengths at different stations. These circles usually intersect at rather acute angles, so even a small difference in measuring (from the strip chart) the signal strength can make a large displacement. This makes for a large imprecision, which they do not estimate.

To summarize, your presumptions of how VAN made their predictions are entirely unfounded. I am quite amazed at how readily you grab these notions seemingly out of the air, and yet (below) you reject the documented "failed to validate". ~ J. Johnson (JJ) (talk) 23:39, 8 October 2016 (UTC) (Slightly belated.)

JJ, back in the dark ages it's true that engineers and scientists used chart recorders, and protractors. Man went to the moon with protractors and slide rules and computers with a few kilobytes of core memory.

Your view that the data was simply "eyeballed" is also unfounded. Obviously the precision of such methods as "apollonian circles" would depend on the care that was used. According to Uyeda, it seems the equipment was upgraded by 1996. And now that it's 2016, it seems pretty silly to me that we're even debating what technology VAN used in 1984. We're wasting our time here, see my comments below about moving forward. JerryRussell (talk) 17:41, 9 October 2016 (UTC) tweaked JerryRussell (talk) 17:54, 9 October 2016 (UTC)

My view is very much "founded", in the articles by Varotsos and Alexopoulos (1984a,b) where they document their methods. You really should read the cited sources.

Jerry, we would waste less time if you would be more focused. E.g., I am not objecting to the use of "dark ages" equipment (I've used a bit of it myself); I am objecting to your presumption that the VAN data is "highly precise", and particularly to your assertion (15:35) that the "squiggley line" I referred to (as metaphor for VAN's raw data) "consists of a series of digital measurements, each of which is typically precise to a couple of digits, and the series can then be averaged, or processed using fourier techniques." You base this on Uyeda's claim (?) that the SES signals were digitized, but you seem to have freely extrapolated beyond that.

We are arguing about what technology VAN has used because you are stuck on this idea (from no where) that VAN's data is "highly precise". Which goes back to the precision (or lack of) in their prediction locations, for which the two telegrams (remember them?) were cited as illustrations.

That any equipment was subsequently upgraded (if it was) is entirely irrelevant, as VAN's 1980's and 1990's claims of prediction are based on what they had then. You can claim that my view is "also unfounded", but you can't argue it on documented basis. I can. So I suggest that the best way to move forward is for you let go of this notion of "highly precise". ~ J. Johnson (JJ) (talk) 21:45, 9 October 2016 (UTC)

JJ, we had been arguing about this very copiously before the topic of the actual precision of VAN's work even came up. It remains my opinion that this particular criticism of VAN is unjustified, and that the examples do nothing to prove it. It's true I haven't read Varotsos 1984, but I seriously doubt that it states that they "eyeballed" the data. So your allegation is no more demonstrated on a documented basis, than my allegation of "plausibility" that their results were as precise as they said.

Nevertheless, the allegation has been made in RS publications that VAN's predictions lack precision, and this information is repeated in the article. And your examples are there, as inadequate as they might be for their alleged purpose. So will you please drop the stick already? Or else make a proposal as to what further changes you would like to make in the article text. JerryRussell (talk) 22:25, 9 October 2016 (UTC)

As I have been pondering this strange discussion, and thinking about my old slide rule that I used in high school, I realized that I always read it BY EYEBALL. And, the Wiki article on the Slide Rule states that they were good to three digits of precision. JerryRussell (talk) 00:19, 10 October 2016 (UTC)

The purpose of citing the telegrams, as stated, is to document the actual wording of a typical prediction, and particularly of their specification of vector direction with an implicit precision of 45° (or 23.5° if you want to be charitable). I think that is important (which we can discuss), but more to the current issue: your sentiment (wish?) that the VAN data is "highly precise" does not appear to supported by any source, and your implication that such data was based on "digital measurements" and Fourier analysis is not only not supported by any source, but also not even likely, given the technology of the time.

I gather you still don't like citing the telegrams, but are there any issues regarding this (other than the general issue of "highly precise") that are not resolved? ~ J. Johnson (JJ) (talk) 22:42, 10 October 2016 (UTC)

As I wrote on 24 Sept., I feel that the article section is too long, and that citing the telegrams is "undue weight". To this day, nothing has changed my opinion about that. But as I also said at the time: if it's important to you to insert the material, I wouldn't edit war about it. I should have added, I wouldn't consider it worth my time to go to any form of dispute resolution about this point, either. As for the ensuing discussion, what can I say? It's been interesting, and we've exchanged a lot of ideas. We aren't in agreement, but I have nothing to add that wouldn't be more or less repeating myself. JerryRussell (talk) 00:07, 11 October 2016 (UTC)

And as I wrote on 26 Sept., previous discussion and editing shows that being too long doesn't seem to apply to addition of VAN-positive (or "balancing") material, and is raised only for deleting criticism, or (as here) of material that illustrates or supports that criticism. While I allow that proper length is a valid consideration, I find that the "too long" objections raised on this page are not proper.

Please note that my original forumulation was much shorter, because I summarized the material. It expanded when the VAN "balancing" material (like "[t]his analysis however was shown by VAN group to be invalid...") was added, and in response to the demand for more citations. As we have discussed.

Lacking any statement otherwise, I will assume that you no longer have an objections to citing the telegrams per se, and particularly not regarding any "selection" criteria.. Time permitting, I will review the text for possible revision, but that is likely to take a week or more. ~ J. Johnson (JJ) (talk) 22:37, 11 October 2016 (UTC)

I've made a proposal below, beginning from your original formulation (for the section on VAN predictions 1987-1995) and showing how I would go about bringing the material into balance, without exploding the level of detail. JerryRussell (talk) 03:17, 12 October 2016 (UTC)

That's fine, but what I am asking here is whether we can close out this section, which is specifically about the use of the telegrams. If you want to discuss "highly precise" let's break that out as a separate section. (We really need to practice staying on topic.) ~ J. Johnson (JJ) (talk) 21:26, 12 October 2016 (UTC)

JJ, what would it mean to you to "close out this section"? Are you asking whether there are any conclusions that can be drawn? If you're planning a revision to the text in the future, I reserve the right to comment about your changes at that time. JerryRussell (talk) 02:05, 13 October 2016 (UTC)

I think we should have some clarification that use of the two telegrams to illustrate the typical wording of the VAN telegrams does not involve any issues of "unjust criteria" or "selection", or any other issue. If not, then let's admit that this has been an colossal waste of time that has not, and is not, going anywhere, and then close it before we waste any more time on it. You don't have to "reserve" any "right" to comment on any future changes, but you should let go of any allegations of "unjust criteria", etc., regarding the use of these two telegrams. ~ J. Johnson (JJ) (talk) 21:33, 13 October 2016 (UTC)

Back on Oct. 5, I wrote: Back on Sept. 28, I wrote: JJ, thank you for clarifying that the 9/1/88 telegram or telegrams were singled out for reproduction in VL 1991, and thus that your selection was not "original research". I'm sorry if that wasn't clear; I've been done with the issue of "unjust" selection, OR, etc. since then. This time, I'm at a loss as to how I could have been any more clear than that, or why you are still berating me for more clarification. JerryRussell (talk) 00:49, 14 October 2016 (UTC)

You said you were "done", but that left open whether the issues are resolved. But I will take your comments as meaning the several issues raised are settled, and that there is no issue. If there are no further comments I will close this section so that some passer-by doesn't get sucked in. Some small progress is made. ~ J. Johnson (JJ) (talk) 22:14, 14 October 2016 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Chopping wood and carrying water for IP202?

Well yes, being helpful to others is a recommended Zen spiritual practice. We've identified a couple of good reasons why IP202 shouldn't be editing the article directly: (1) he's an SPA here for advocacy purposes, and (2) English is not his first language. So if he comes to talk and makes edit requests, and I evaluate those requests and fulfill them if I feel they comply with policies, I don't see the problem. If we're not reasonably responsive to IP202 requests, the gentlemen's agreement breaks down, and IP202 might pursue other dispute resolution paths.

Also, in accordance with WP:BRD, I don't see any problem even making bold edits first, and then discussing them afterwards on the talk page, perhaps after those edits have been reverted. In the particular example we've been discussing, I did in all good faith believe that the matter had been discussed first. But even if that had not been the case, WP:BOLD is the operative policy here. Of course the ultimate outcome of the process depends on achieving consensus. JerryRussell (talk) 18:26, 27 September 2016 (UTC)

So when an identified SPA requests some edits you feel quite comfortable in evaluating those requests entirely on your own, without any discussion (let alone concurrence) with anyone else? Because if you are not responsive the SPA might violate some kind of "gentlemen's agreement"? Would that supposed agreement cover civility, COI editing, or promotion?

I can't imagine why you would think there had been a sufficiency of discussion (let alone that consensus had been achieved) on any of this SPA's edit requests. But as you are so congenial with WP:BRD I am wondering if the "R" part should be exercised across all of your recent edits. And then we can "D" some more. ~ J. Johnson (JJ) (talk) 21:26, 28 September 2016 (UTC)

Please consider the bold edits as proposals. I've tried to make my proposals consistent with policies on COI and promotion. In hopes of saving us both some work, I've taken the controversial texts and moved them into the footnotes for the time being. A few minor changes have not been rolled back in this way, but I hope you'll find them non-controversial. I trust you'll take whatever "R" steps you feel are necessary to protect the integrity of the article, and "D" will continue. JerryRussell (talk) 22:26, 28 September 2016 (UTC)

After all the hacking that's gone on the last four months I am inclined to revert everything and start afresh. To get all the horse manure out is going to take a broad brush, not a small brush. But we can discuss it. Note that "proposals" are best presented as proposals, and not just slipped in and see if anyone notices. ~ J. Johnson (JJ) (talk) 23:51, 29 September 2016 (UTC)

I've just learned about the policy wp:preserve, which recommends to "Preserve appropriate content". This policy gives many alternatives to blanket content removal. JJ: The existing "horse manure" embodies the results of a recent RfC in which, in many cases, you were a dissenting voice in opposition to the consensus. Please keep that in mind if you are contemplating massive reversions of existing content. Of course the same considerations would not apply to my bold edits from the last few days.

It is basically impossible at Wiki to "slip in" editing proposals. Everything is quickly revealed by doing "diffs", if you have concerns that you might not have noticed something has changed. JerryRussell (talk) 15:28, 30 September 2016 (UTC)

You are thinking in terms of text "magically" appearing without any record, which I quite understand, thank you. What I was referring to is all of the little additions and deletions and modifications, where examination of individual edits doesn't necessarily reveal the overall trend and result. ~ J. Johnson (JJ) (talk) 00:26, 5 October 2016 (UTC)

I can understand your concern. I made about 20 edits between the 25th and 28th of Sept., with Bender the Bot slipping a few in along the way. It's possible to use the radio buttons and "compare selected revision" button to generate diffs across a sequence of edits. JerryRussell (talk) 04:04, 5 October 2016 (UTC)

Mileti & Sorensen

JJ, I just noticed that you also reverted my citation to Mileti & Sorensen. Did you mean to do that? JerryRussell (talk) 02:39, 25 September 2016 (UTC)

That passage got deleted because I didn't feel like taking any extra trouble to tease that edit it apart from the other material you removed. Perhaps you would have taken that extra trouble, but you didn't step up to the task so I did it my way. You could restore the M&S passage, but I think you really shouldn't. You got M&S from IP202's rather confused comment (05:54, 16 Sep.) that the article was "making a wrong impression on alarming}", which he seemes to think demonstrates "Orignal Research" in the existing table. From his reference you took some concern (above, at #Existing contingency table) that "the public simply does not panic ...", or that "the "cry wolf" syndrome is greatly overrated", or something. (You never did explain the point of your concern.) As near as I can make out, the only connection of all of that with the table is in the single word "panic". A very tangential item.

In short, you have not shown any problem for which M&S is a necessary solution. I suggest leaving it out. ~ J. Johnson (JJ) (talk) 23:49, 25 September 2016 (UTC)

I'm sorry you're having difficulty understanding "the point". The problem with all the precursor methods is a high false alarm rate. Nobody disputes this, although there's an ongoing debate about whether any of the precursors are associated with earthquakes at all. If we grant that there is some correlation between precursors and EQ's, the question is: can we do anything with predictions that have high false alarm rates?

The big problems with false alarms are: (1) Panic, and (2) "cry wolf" effect.

M&S suggest that panic & "cry wolf" are both overrated, and can be mitigated by proper communication of the alarm to the public. That's important for our readers to know. JerryRussell (talk) 01:55, 26 September 2016 (UTC)

You are flat out wrong. The problem I have had with your "point" of objection is that you not stated it. So what is it?

1) That the "problem with all the precursor methods is a high false alarm rate"? Please note: that statement is FALSE. In general, the false alarm rate of any predictor ("precursor" or not) can be forced very low, and even to zero, very easily: just raise the alarm threshold ("The Bar"). Of course, you miss all of the hits, but that is the nature of the trade-offs between Type I and Type II errors.

2) Or that "The big problems with false alarms are: (1) Panic, and (2) "cry wolf" effect."? Again FALSE. There is no statement in the article (in any version), nor any basis for your statement, that "The big problems with false alarms are: (1) Panic, and (2) "cry wolf" effect." The big problem with false alarms is economic disruption. If the current text is not clear enough on that perhaps we should restore some previously cut text.

3) Or that panic and "cry wolf" "can be mitigated by proper communication of the alarm to the public."? This appears to be true, but it is quite irrelevant. That the possibility of panic is reduced might alter the trade-offs, but does not remove the dilemma: You still can't reduce one without increasing the other.

In the section where you raised this issue (#Existing contingency table) I thought I had made it emphatically clear that the table is about the DILEMMA of issuing an alarm, or not issuing alarm. (Excuse me for asking, but is it possible you don't know what a dilemma is?) To state this in terms of the picture: it's NOT about the rock, and it is NOT about the whirlpool. It is about trying to steer between them. Having M&S tell us that the whirlpool might not be quite as deep as it looks does not make the steering any easier. (And if they were in my boat I'd tell them to shut up and sit down.) ~ J. Johnson (JJ) (talk) 21:56, 26 September 2016 (UTC)

Yes, I know what a "dilemma" is, and I do not claim the dilemma disappears. It can, however, be mitigated. The current text mentions economic losses as well as panic and 'cry wolf' effect, and it's your judgement (not any statement in the sources) that economic damages are the most important. In my opinion, the billionaires who are suffering the bulk of economic losses, are not more important than the people in substandard structures being killed, but that's just me. JerryRussell (talk) 22:08, 26 September 2016 (UTC)

I forgot to mention: with respect to: In general, the false alarm rate of any predictor ("precursor" or not) can be forced very low, and even to zero, very easily: just raise the alarm threshold, of course I see this and I agree. But if the alarm threshold is set so that a useful "hit rate" is obtained, then the false alarm rate will be high, at least for the real-world methods we're discussing. JerryRussell (talk) 18:29, 27 September 2016 (UTC)

That's right, that's just you, and the intrusion of your personal feelings into the article (including our editorial process). But your butterfly-like concern that (how does this go?) M&S should be cited (and not just in a note, but named right in the main text) lest a sub-optimal communication of an alarm should result in the death of someone in a substandard structure? That is so distant from the key point of that section you might as well mention kittens and puppies (always a crowd-pleaser). Now I wonder if you have noticed where someone said that earthquakes don't kill people, failing structures kill people. It seems to me that some lives might be saved if that was better communicated. As it is, I've been wanting to write something on that since last spring, but it seems all my spare time goes to the damn penny-ante quibbling on this article. At any rate, your concern that "panic" and "cry wolf" were dissed in the table in favor "economic disruption" is unsourced, and not even relevant to the concept demonstrated. It is entirely a distraction. Neither the table nor the text (nor even myself) say anything about what is "most important". And citing M&S adds absolutely nothing to the article.

At last you seem to grasp that there is a TRADE-OFF between a useful hit rate and an acceptable false alarm rate. But whether the trade-off should be set to favor billionaires or kittens, and how to do so, is quite outside the scope of the article. M&S might be relevant in an article on the communication of alarms. But if this article were to uniformly cover that kind of detail we would have a multi-volume work. ~ J. Johnson (JJ) (talk) 21:33, 28 September 2016 (UTC)

The section certainly does mention death and destruction as consequences of our lack of ability to predict earthquakes. I hardly know what to say in response to your trivialization of human death as somehow equivalent to kittens and puppies. If the topic of this section is "how to set the bar", then lowering the bar results in more hits (which should translate to lives saved) as well as more false alarms. So if the cost of false alarms is over-estimated, then lowering the bar saves lives. So the M&S quote still seems highly relevant to me.

I wonder if you have noticed where someone said that earthquakes don't kill people, failing structures kill people. Have you noticed that the very next paragraph, that exact point is clearly spelled out?

The topic of the section is NOT "how to set the bar". (Bolded to aid your attention.) The point the contingency table addresses is that in trying to set the bar – where ever one sets it, and for what ever ends you want – there is a DILEMMA. That I don't immediately join you in demanding that the bar should be set for the maximal preservation of human life is NOT "trivialization". It is also ludicrous, as a maximal setting would be: expect a major earthquake ALL the time. And NEVER enter a building that might collapse. Or cross a bridge. Or walk past URM buildings. Etc. That would save lives, until people noticed your predictions were no damn good, and just ignored them. ~ J. Johnson (JJ) (talk) 06:18, 2 October 2016 (UTC)

JJ, isn't the topic of the section something that we should determine collaboratively, by consensus, as editors of the article? The existing article contains the sentence The acceptable trade-off between missed quakes and false alarms depends on the societal valuation of these outcomes, which I would interpret as addressing the question of "how to set the bar". The text within the dilemma table states "lowering the bar reduces odds of losses, but increases the costs of false alarms." Again, this seems to qualitatively address the question "how to set the bar".

I agree it would be ludicrous to argue that the bar should be set so low as to expect a major earthquake all the time. To use a concrete example, the most optimistic estimate of VAN 1990's capability is that it could achieve a 70% hit rate at the cost of a 90% false alarm rate. If that's the case, perhaps the alarm could be communicated to the public as follows: "Some seismologists believe there's a 10% chance of a major earthquake within the next two months, somewhere within 120 km of location X." I doubt if such a prediction would elicit much panic. I'm not sure how useful it would be in terms of saving lives, either; but it seems possible that such a warning might prompt some people to take specific precautions.

It has also been complained that VAN predictions were active much of the time. VAN made 94 predictions over an 8 year period, each one of them open for a month or more. So it does seem that on the average one of those predictions would always have been active. Wiki gives the surface area of Greece as 131,957 square kilometers. If the effective radius of VAN predictions is 120 km, that is 45,216 square kilometers, or about 1/3 the surface area of Greece. So on average, about 1/3 of the area of Greece would have been covered by an active VAN prediction at any given time. (This rough calculation does not account for VAN predictions centered in the oceans.) I wouldn't encourage using this calculation in the article, as it goes beyond simple arithmetic in its presumptions, which makes it WP:OR. But it tends to indicate VAN predictions would indeed have been active often enough that "cry wolf" would be a rather difficult problem to mitigate; while at the same time, any claim that VAN predictions "expect a major EQ all the time" is going beyond the facts. JerryRussell (talk) 17:14, 2 October 2016 (UTC)

You want a different topic at that point in the article? Then please make a suggestion. Although it seems to me that currently we have way to many unresolved issues to be starting any new ones. In the mean while, please note that the evaluation of a prediction is quite an ample topic itself, without extending it to the communication of alarms. M&S touches on the latter, not the former. ~ J. Johnson (JJ) (talk) 00:36, 5 October 2016 (UTC)

I don't think I'm asking for a different topic. I think the topic of the existing section can be construed more broadly than you're suggesting. But enough of that. You're right that we've got too many issues open right now anyhow.

And, you've actually won me over to your point of view about the Mileti & Sorensen quote. On some abstract level, the concept of alarm communication seems important. But considering that there's really no proof that any of these short-term prediction methods work well enough to ever warrant any sort of alarms, maybe it's WP:UNDUE to call attention to M&S. It's OK with me to just leave it in the footnote, if that's OK with you. JerryRussell (talk) 04:15, 5 October 2016 (UTC)

Okay! I note that to the extent society wants alarms to be useful the communication of the alarm is important. And the evaluation of a prediction certainly figures in to what what should be communicated. But it doesn't work the other way. That is, how an alarm is communicated does not affect the evaluation of a prediction. It might very slightly affect the decision criterion of whether to issue an alarm, but keep in mind that the text is not about when, or even how, an alarm is issued; it's about the trade-off, that you can't "improve" the prediction on one dimension without worsening it on the other. And I would say that M&S is so far from this there is no reason to have it all, not even the footnote. It is not really directly relevant to anything here. ~ J. Johnson (JJ) (talk) 20:35, 5 October 2016 (UTC)

Proposed additional text re: VAN SES method

The article up until recently has read as follows:

Objections have been raised that the physics of the VAN method is not possible and the analysis of the propagation properties of SES in the Earth’s crust claimed that it would have been impossible for signals with the amplitude reported by VAN to have been transmitted over the several hundred kilometers distances from the epicenter to the monitoring station.[58][59] It was also claimed that VAN’s publications do not account for (i.e. identify and eliminate) possible sources of electromagnetic interference (EMI).[60]

At IP202's request, I proposed adding the following information:

However, the VAN group claims that SES signals mainly travel through fault zones whose conductivity exceeds significantly, i.e., by a factor 102 to 103, the conductivity of the surrounding medium.[59] They also claim to have derived and published criteria that can clearly distinguish SES from other electric signals of no precursory nature.[57]

Discussion question: May this additional information be included in the paragraph? (It is currently in the footnotes.) Or must it also be deleted from the notes as well? JerryRussell (talk) 23:12, 28 September 2016 (UTC)

comment I believe the additional information must be included, to avoid giving a false impression that VAN has failed to consider the issues raised. The existing content is correct in itself (the sources did make these claims of VAN errors) but for us to make an editorial decision that these remarks should be given without also providing equal prominence to VAN's response, is a violation of WP:NPOV. Furthermore it can be construed as creating an intentional false impression, and thus derogatory to VAN. JerryRussell (talk) 15:38, 30 September 2016 (UTC)

Physics of SES propagation is knowledge, given simply, especially since the critique on this subject is considered notable, and thus is Wikipedia article material, not part of the footnotes. It remains to be decided which article should contain both. The criteria are also physics but cannot be explained simply. We could move their explanation to the footnotes, if needed, but still there should be a visible hint that there is a physics model answering the critique. The impossibility of the physics of VAN method should not stay in the article, unbalanced.--IP202-77.69.68.33 (talk) 16:17, 30 September 2016 (UTC)

Proper application of WP:NPOV and WP:FALSEBALANCE with respect to VAN

WP:FALSEBALANCE gives examples such as flat-earth and moon landing conspiracy theory, which are clearly deprecated by the mainstream. Is VAN in that same extreme category of pseudoscience? I see no evidence that this is a dominant mainstream position. In fact there is a diversity of views about VAN.

I hope that everyone can agree that for 1990's VAN technology, if the alarm "bar" was set so as to achieve a reasonable hit rate (~70%), the false alarm rate was very high (~90%). The predictions also lacked spatial and temporal precision. So that created a very real limit on the utility of the predictions. VAN has worked to improve that problem with "natural time" but the jury is still out, whether there has been any success on that front.

Some sources go beyond that, and claim that VAN method is no better than random darts thrown at a calendar, the hit rate was much closer to zero, and that VAN are charlatans. But I don't see evidence that this is The mainstream position. In my opinion, writing our article as if this perspective is WP:TRUTH is inappropriate. An example case in point is the discussion directly above, on "Proposed additional text re: VAN SES method". What possible reason would there be to oppose inclusion of the extra text, except for an inappropriate application of WP:FALSEBALANCE? JJ, I get that you're opposed to including this material, but I confess I am resorting to mindreading here to try to come up with why. Am I understanding you correctly? JerryRussell (talk) 01:51, 1 October 2016 (UTC)

Jerry: all that you have opined about here arises from your refusal to accept the mainstream opinion that VAN is not accepted science. But consider: if VAN was right, then why isn't that science used to predict earthquakes? Varotsos and Uyeda say it is because they are not properly funded, but as has been established before both have been amply funded, for over twenty years.

As to evidence (documentation) that VAN is NOT accepted science I again refer you to the reports of the ICEF, prepared by a team of of internationally recognized experts from the U.S., China, Italy, France, the U.K., Greece, Russia, Japan, and Germany. From the executive summary (2009), §A:

Given the current state of scientific knowledge, individual large earthquakes cannot be reliably predicted in future intervals of years or less. In other words, reliable and skillful deterministic earthquake prediction is not yet possible.

Or §C:

The Commission has identified no method for the short-term prediction of large earthquakes that has been demonstrated to be both reliable and skillful.

From the final report (2011), p318:

... large earthquakes cannot be reliably predicted for specific regions over time scales less than decades.

Or p319:

... the inability to reliably predict large earthquakes in seismically active regions on short time scales. ... The search for diagnostic precursors has not yet produced a successful short-term prediction scheme.

Those were general statements, referring to ALL methods of prediction. Where electromagnetic signals are discussed as possible precursors VAN's claims are specifically addressed (p335):

... subsequent testing has failed to validate the optimistic SES prediction capability claimed by the authors" [emphasis added].

In the end it is very simple: VAN claims an ability to reliably predict earthquakes, the ICEF says: no, VAN's claims failed to validate. (Can you see this?) You say it is inappropriate to write the article from that perspective, but why do you insist the article should be written from VAN's perspective? Because VAN keeps turning out crappy papers which the VAN enthusiasts here give more credence than all of the mainstream papers? ~ J. Johnson (JJ) (talk) 23:13, 3 October 2016 (UTC)

Hi JJ, I disagree that I'm asking for the article to be written from VAN's perspective. I'm recommending a neutral perspective.

If you look at the ICEF report, when it comes to the statement failed to validate, the reference given is [151] Geller [1996], Lighthill [1996], Mulargia and Gasperini [1992]. So we're back to the same sources we've been discussing, and the ICEF's one-paragraph summary leaves out a lot of nuance. And, their summary asks as many questions as it answers. What are the "optimistic" prediction capabilities claimed by VAN? Does this mean the breathless claims in the first few papers they published, which certainly have failed to materialize? Or the more sober results presented in Lighthill, where they are basically admitting to a 90% false alarm rate?

Suppose VAN's claims are true, that they have a prediction method with a 90% false alarm rate. How would that debunk all those other quotes from experts saying that EQ's cannot be "reliably predicted"? It all seems to paint a consistent picture to me.

Perhaps your summary dismissal of VAN's more recent "crappy" papers is betraying some bias on your part? I mean, those "crappy" papers are being published by the most prestigious academic peer reviewed journals. JerryRussell (talk) 00:00, 4 October 2016 (UTC)

And just what is wrong with the "same sources we've been discussing"? Is your argument that 1) we can't use anything from Geller (despite him being one of the foremost experts on EP, and his review the most comprehensive ever done) because a) he is antipathetic to VAN, and b) his h-index is less than Varotsos'? 2) We can't use anything from Lighthill because it includes stuff from Geller? 3) We can't use M&G 1992 because that is old science? Therefore, 4) we can't use the ICEF because they rely on all of the foregoing? Is that how it goes?

That argument is pathetic. And I am not much impressed with your question, "What are the "optimistic" prediction capabilities claimed by VAN?". Yes, perhaps those breathless claims in their "first few papers", which would include Mary Lazaridou's 2012 book about "The success of the VAN method over thirty years". So why this quibbling about the ICEF? You were wanting a definite verdict from some kind of scientific jury re VAN, but when you get it you complain you don't like their references.

I don't know what you mean by "most prestigious academic peer reviewed journals", but VAN's papers are NOT being published by the most prestigious scientific peer reviewed journals (Nature and Science), nor even in leading earth science journals such as the Journal of Geophysical Research, the Geological Society of America Bulletin, the Bulletin of the Seismological Society of America, and Geophysical Research Letters.

And when I say "crappy" I mean for some of the same reasons you have acknowledged (such as lack of precision). ~ J. Johnson (JJ) (talk) 21:17, 5 October 2016 (UTC)

No, I'm not arguing that we can't use anything from Geller, or Lighthill. I'm arguing that the GRL and Lighthill volumes include articles by Varotsos and his supporters as well as Geller and other VAN critics, and it's not appropriate to quote only the critics. The ICEF summary is brief, ambiguous, and adds nothing to the debate.

I have stricken my superlative "the most prestigious..." from my description of VAN publications. I'm sure the editors of journals such as Physical Review, Applied Physics Letters, Tectonophyics, and PNAS will be deeply concerned that certain Wikipedia editors think their publications are "crappy" because of some arbitrary notions about notational precision. And there's obviously no prestige at all in getting published by Springer-Verlag. JerryRussell (talk) 00:29, 6 October 2016 (UTC)

I am so glad to hear that is not your argument. But then, what is your argument? That the "ICEF summary ... adds nothing to the debate"? I think you have confused the ICEF's assessment of the debate – the verdict of the scientific jury, if you wish – with the debate itself. They weren't "another critic", they were the jury.

And you will please note that the original version of these sections only summarized the findings. The critics were quoted when a certain editor started demanding sources. Adding balancing quotes gives the impression that these issues are still in controversy (they are not), or that VAN's pov is equally valid (it is not).

Jerry, please don't let snarkiness (okay with me) lead you to misstatement (bad!!). Even if one allows that some VAN articles are crappy, and even if that is taken to reflect on the editors (but note that editors often good reasons for publishing articles with distinct shortcomings), it is quite a leap to extend that to the entire publication. Although I will say that PNAS is notably weak in earth sciences, which is why Pullinets tends to publish there than in the mainline journals. Also, these "notions" about precision are NOT arbitrary; do find an introductory text to review if you won't take my word for it. ~ J. Johnson (JJ) (talk) 21:27, 6 October 2016 (UTC)

Inside ICEF it is clearly stated that Papadopoulos alone is responsible for Greece sayings. He said no word for 2008 earthquake or natural time, but the publication on these was left (mis-cleanup?) within the references. This publication is a collection of POV reviews and this is undeniable. "Physical Review Letters" is top and scientists focusing on critique as a career do not publish there, as only really important publications have place there. H-index is not of linear scale. Natural time is also forgotten here as an extra criterion of verification of SES nature with impressive results and an increase of time precision of a VAN prediction (or any critical phenomenon). We are stuck to decades ago on arguing here, but even so there is still no response on the (also old) SES propagation issue above--IP202-77.69.97.224 (talk) 08:40, 7 October 2016 (UTC).

All nonsense. The ICEF report is not a "collection of POV reviews", it is a review, by some of the world's top experts. You object to the SES because their views are disfavorable to VAN, but as we have seen you are entirely a VAN partisan. You try to trivialize VAN-critical views by suggesting they are due to only one or two individuals (first Geller, now Papadopoulos), ignoring the broad basis of such criticism. You complain we are "stuck to decades ago" arguments, but you reject the most recent assessment of those arguments.

That Physical Review Letters is "top" in physics is irrelevant, as the field here is seismology. ~ J. Johnson (JJ) (talk) 22:01, 7 October 2016 (UTC)

JJ, it's not me whose snarkiness is leading to misstatements. We all know that PNAS is so weak in earth science that they run articles about flat-earth and creationism on a regular basis. It's so bad that they even accept "crappy" articles by Pullinets. OMG.

But, flat-earth theory and Creationism get more of a fair presentation at Wikipedia than you're willing to allow for VAN. JerryRussell (talk) 19:13, 7 October 2016 (UTC)

Really? So this is the sort of thing you'd like here or over in the VAN Method article?

The overwhelming consensus of the scientific community is that creation science is a religious, not a scientific view. It fails to qualify as a science because it lacks empirical support, supplies no tentative hypotheses, and resolves to describe natural history in terms of scientifically untestable supernatural causes. Creation science is a pseudoscientific attempt to map the Bible into scientific facts, and is viewed by professional scientists as unscholarly and even as a dishonest and misguided sham, with extremely harmful educational consequences.

I mean, I suppose it's fair in the sense that it's accurate, but in what sense is the treatment of VAN less fair? --tronvillain (talk) 22:08, 7 October 2016 (UTC)

Jerry: Really? Does "neutrality" require (or even permit) every assertion and implication that the earth is round be "balanced" with a response from the Flat Earthers? Where is the Round Earth POV being "balanced" by Flat Earth partisans? ~ J. Johnson (JJ) (talk) 22:13, 7 October 2016 (UTC)

This article includes all the equivalent denunciations of VAN: that EQ prediction may be impossible, that there is no physical mechanism that could account for any positive results, that their results are no better than random chance, their program is a huge waste of money, they cause panic and never have saved any lives. All of those negative views are clearly stated, and rightly so. It may all be true.

But in the creationism article and the Flat Earth article, it's possible to read what the partisans of those points of view have to say. Here, VAN's opinion about the conduction mechanism is compressed into a footnote. JerryRussell (talk) 02:41, 8 October 2016 (UTC)

In an article about a fringe view of course we describe that view, and even why the partisans have adhered to it. And in VAN method there would be a lot more scope to describe VAN's views (but NOT to give any suggestion of "equal validity"). But in case you have forgotten, this article is about earthquake prediction generally, of which VAN is just one painful episode of a failed method. A failed (invalid) method is not owed any rebuttal.

So please answer my question: Where is the Round Earth POV being "balanced" by Flat Earth partisans? ~ J. Johnson (JJ) (talk) 23:36, 8 October 2016 (UTC)

Could it be fair to intentionally discard all the published results of an earthquake prediction method in well known peer reviewed journals like Physical Review, PNAS, Tectonophysics, Journal of Geophysical Research etc during the recent decade, also documented in advance in the official site www.arXiv.org, and now claim (as JJ did) that it is "a failed (invalid) method [which] is not owed any rebuttal" ? Could a reader or a wikipedian or an administrator ever imagine that an author of a wiki-article has proceeded to such unusual actions?--IP202-178.59.54.176 (talk) 08:41, 9 October 2016 (UTC)

JJ: as it was determined in the RfC, this is an article about earthquake prediction, which is generally a fringe topic in which all methods have failed. Therefore, it is appropriate to treat VAN in a balanced fashion within this article, which would include accurate presentation of their point of view.

In case you missed my sarcasm here, I haven't actually agreed that VAN is in the same category as flat-earth theory. The problem is with the distinction between EQ prediction ("reliable & skillful alarms, suitable for evacuation of cities") vs. EQ forecasting. "Operational Earthquake Forecasting" is not fringe, and it's far from obvious to me that VAN has failed as a short-term forecasting method. JerryRussell (talk) 15:18, 9 October 2016 (UTC)

It may be a half-filled or half-empty glass. Publications deal with water in the glass. This is what our article should present too, after saying we have not reached the top of the glass. We do not care in Wikipedia what the truth is about earthquake prediction, beyond that it is not achieved. We care here for what is notable and verifiable by the sources.--IP202-178.59.54.176 (talk) 16:39, 9 October 2016 (UTC)

I want to add (since JJ is still emphasizing the ICEF report's "verdict" above) that: (1) The ICEF comments are in the context of "reliable and skillful" EQ prediction, which VAN is admittedly not; (2) the 'verdict' is ambiguous, as it's not clear what optimistic standards have not been met; and (3) the ICEF was a highly politicized team organized for the defense of the Italian scientists who stood accused of manslaughter, as much as it was an objective scientific committee. Therefore I reject the idea that their report is the last word on the topic of VAN's success or failure. JerryRussell (talk) 17:49, 9 October 2016 (UTC)

Of course VAN isn't in the same category of the flat-earth "theory": proponents of the latter may have never existed (see Myth of the flat Earth), are certainly not alive now, and don't have reputations and massive funding at stake. Most certainly they don't have partisans still beating the drums on their behalf. But VAN is still, per the mainstream, a failed method.

The ICEF's verdict is not ambiguous, it was clearly: testing failed to validate.

VAN's "optimistic claims" are (as stated in the article) that:

there was a "one-to-one correspondence" between SES and earthquakes", – that is, that "every sizable EQ is preceded by an SES and inversely every SES is always followed by an EQ the magnitude and the epicenter of which can be reliably predicted" – the SES appearing between six and 115 hours before the earthquake.

Seems quite clear to me.

Your assertion that the ICEF was "a highly politicized team organized for the defense of the Italian scientists" is unsourced, and sounds like yet more stuff that you just pluck out of thin air. They were asked to "assess scientific knowledge of earthquake predictability" (see the Preface, and also the Abstract), and I see absolutely no showing their verdict is incorrect, or even contested.

Whether the ICEF was the last word on VAN is beside the point. The ICEF is the latest scientific assessment from the mainstream. Your persistent quibbling about this displays a strong non-neutral point of view. ~ J. Johnson (JJ) (talk) 23:29, 9 October 2016 (UTC)

References

JJ: My opinion of ICEF is purely my own opinion. It is our role as Wiki editors, to assess the quality of the sources. VAN has been discussed in many, many sources since the 1996 papers cited by ICEF. If you think my POV is non-neutral, so take me to AN/I. I dare you. JerryRussell (talk) 23:55, 9 October 2016 (UTC)

And if you aren't going to take this to AN/I, here and now, I demand a retraction of your statement that I have a strong non-neutral point of view. I construe it as a personal attack. JerryRussell (talk) 00:05, 10 October 2016 (UTC)

Well JJ, you finally got to me, and you made me angry. But on reflection, I'm not sure whether AN/I would be the best venue for dispute resolution. So I'm retracting my ill-considered remarks above. If you aren't enthusiastic for another RfC, maybe we could try DRN, or mediation. Or if you want to take it to AN/I, I'm game for that too, whatever.

ICEF passed on a great opportunity to actually read & review the many papers produced by VAN and their friends since 1996. How pathetic, and what a missed opportunity, that they referred to the 1996 papers and left it at that. We've already talked about the hatchet job they did on Giuliani.

ICEF didn't say whether they were judging VAN by the criteria of their 1984 claim, but if they did, so much the better. I believe VAN has significantly qualified those claims since 1984, though obviously they still believe there is some sort of correlation between SES and EQ.

I don't agree that our problem is any lack of neutrality on my part, but I do agree that this endless quibbling is getting us nowhere. JerryRussell (talk) 07:56, 10 October 2016 (UTC)

Jerry, I am truly sorry if you are angry, or feel provoked. And if "strong" was too much I willingly strike it. (Done.) But it does seem to me there is a certain partiality in your point of view. Please do not take it as a personal attack that I raise this. I see it as an error (and don't we all err?) which can be corrected, not any kind of personal trait. That you are persistent in that view could be taken as a personal criticism, but that really depends on why. E.g., I am persistent in my view, but I think that is warranted by facts, cogent argument, etc., not because I am (say) influenced by tabloid journalism. Do you not also think so for yourself? Then why can't we resolve all of this on a rational, and even collegial, basis?

As to the issue at hand (the ICEF and their assessment), I am still troubled by your comments. E.g., why is it "pathetic" that the ICEF "referred to the 1996 papers and left it at that"? To the extent all of the 1996 discussion settled the matter for mainstream seismology, is there anything that VAN (and their friends) have said, or even could say, that changes the mainstream point of view? How does "pathetic" even enter into this? ~ J. Johnson (JJ) (talk) 23:09, 10 October 2016 (UTC)

Thanks for the offer of ongoing collegial discussion. I'm sorry I got triggered.

I don't question that your view is warranted by facts, cogent argument, etc. But, a very diverse and mutually contradictory range of ideas can similarly be backed by facts and cogent argument. The Wikipedia ideal isn't so much that we would all agree on a set of facts based on cogent argument and sources. At Wiki we have a much more modest goal, that all the various points of view get represented neutrally, according their weight in reliable sources. Inasmuch as the goal is to encompass all human knowledge, and there is a "demarcation problem" regarding which fringe views might ultimately win greater acceptance, Wiki generally tolerates articles about fringe topics, as long as there's no false equivalence.

There is room for nuance about what the "mainstream" position really is, and how the ICEF relates to that mainstream. In the 1996 publications, especially Lighthill, there was obviously a diversity of opinion represented. It wasn't as if VAN was alone in arguing for their own position. So now it's 20 years later, and VAN has gone right on getting funding, diligently publishing in excellent peer reviewed journals, and claiming significant innovations. So the "mainstream" is too proud to even mention any of that exists? Yes, I think that's pathetic.

You take exception to my remarks that ICEF is politically motivated, but isn't it obvious? They were chartered by the Italian Civil Defense Dept. (DPC) while their scientific staff was being prosecuted for manslaughter. You don't see any COI in that situation? Really? JerryRussell (talk) 00:39, 11 October 2016 (UTC)

I would say that your statements demonstrate a biased interpretation of the authority you allow these different viewpoints. E.g., you don't allow that the DPC may have been sincerely trying to get an honest scientific assessment, or that the several NON-ITALIAN experts aren't really puppets of the DPC. Yet you would accept on full authority the non peer-reviewed book by Varotsos' wife, as if she would never be influenced by her husband. You also embellish the publications of VAN, which have not made it into the most "excellent" peer-reviewed publications - and why can't you acknowledge that?

A fuller, deeper consideration of your comments (here and below) and expressed POV is deferred, as I don't have that much time at the moment. ~ J. Johnson (JJ) (talk) 23:33, 11 October 2016 (UTC)

JJ, I don't question that at least some of the various scientists on the ICEF might have been seeking an honest scientific assessment. In fact I believe the most relevant conclusion regarding our topic is almost certainly correct: that there is no short-term EQ prediction method that is sufficiently reliable and skillful to justify evacuation of cities. However, I also feel that to some extent the proof of the ICEF's work is in the pudding. We examined the treatments of Giuliani and of VAN, and IMO the comments on Giuliani amounted to a hatchet job, while the failure to even cite any VAN work since 1996 is negligent at best. The situation with the DPC represents a possible COI, and our spot check leads me to believe that this potential COI led to very real problems with the report.

Please note that I refrained from using the superlative "most" with respect to the excellent peer reviewed journals where VAN publishes their materials. I am admitting that in the academic pecking order, there are even more highly esteemed journals. IP202 has discussed the relative impact factors of VAN and their critics, but I'm not really interested in getting into that level of analysis.

Academic books are generally not put through "peer review", but the better presses do have an editorial process. A knowledgeable editor is assigned, and these presses do not publish books by known charlatans in order to boost sales figures. So the book by Varotsos' wife does meet minimal Wikipedia standards for reliable source publications, although obviously it is not what I would call an "independent" source. Certainly it is an adequate source for information about what VAN (as a fringe group) says about their own work. JerryRussell (talk) 00:33, 12 October 2016 (UTC)

I think the mainstream scientific view is that "there is no short-term EQ prediction method that is reliable and skillful", period. Certainly not sufficiently to justify evacuation of cities, but not at any level. (Presumably we both share an understanding of what "skillful" means in the context of prediction.)

That MLV's book meets the minimal Wikipedia standards is quite beside the point. She is not an internationally recognized expert, has weak publication record (off-hand I don't recall that she has published anything on seismology but that she co-authored with Vartosos), and it does not appear that she had any science to add. At best her book is a historical record from the VAN perspective, and possibly useful as an index to their work. On the otherhand, the ICEF is multiple, acknowledged experts, some, including the chairman, quite remote from the DPC. I am not aware of any source (are you?) that questions their independence. You accuse them of negligence for not citing "any VAN work since 1996", but 1) I don't see that you have the competence to make that judgment, 2) what was said up to 1996 was sufficient to show a failed validation, and 3) I suspect the ICEF felt that nothing VAN has said afterwards shows any reason to change that.

I would point out that if you would like me to provide considered (and hopefully articulate) responses to your queries and comments, instead of (like the general run of WP editors) just bandying about the first thoughts that come to mind, you really need to back off from this constant volleying of comments and let me have some time to ponder. I don't believe this should be race.

JJ, how could we share an opinion about what "reliable and skillful" means? In searching the ICEF report that seems to have coined this particular evaluative notion, there is no quantitative definition. It seems to mean pretty much whatever anybody might think it means.

Why didn't the ICEF mention any of VAN's work since 1996? It might be for the reasons you suggest, but then again it might be because they have no explanation for the apparent success of the VAN 2008 prediction, and decided it was safer to sweep that under the rug. Since they didn't say, your guess is as good as mine. And it's our job as Wiki editors to evaluate the sources, as unqualified as we might be for that job.

I've been doing my best to keep up with the constant flow of comments from you, and would welcome it if you take some time to ponder. This is a voluntary activity, participation is not required. JerryRussell (talk) 02:21, 13 October 2016 (UTC)