Wikipedia:Reference desk/Archives/Mathematics/2013 September 6

Welcome to the Wikipedia Mathematics Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

September 6

numeral one

What is the name of the rule of thumb/theory that, in a given set of numerals, the numeral one should appear significantly more often than any other numeral? I have heard that this "rule" is sometimes used to detect whether data has been faked, for example in polling station returns and such. I looked through our article on 1 (number), but didn't see anything. Matt Deres (talk) 03:24, 6 September 2013 (UTC)

You're probably thinking of Benford's law.Phoenixia1177 (talk) 03:52, 6 September 2013 (UTC)

That's it. Thanks! Matt Deres (talk) 10:45, 6 September 2013 (UTC)

1 (number)#Mathematics says: "One is the most common leading digit in many sets of data, a consequence of Benford's law." PrimeHunter (talk) 12:34, 6 September 2013 (UTC)

Limits of Fresnel integrals without complex analysis

See Talk:Fresnel integral#Computation of limits without complex analysis.--Jasper Deng (talk) 04:02, 6 September 2013 (UTC)

I'm still seeking answers here.--Jasper Deng (talk) 00:27, 9 September 2013 (UTC)

How can I refute the mathematical theorem?

How can I refute the mathematical theorem — Preceding unsigned comment added by 37.238.108.105 (talk) 11:41, 6 September 2013 (UTC)

A mathematical theorem is refuted when someone manages to break it, so the short answer is that you need to show how you broke it. Roger (Dodger67) (talk) 11:49, 6 September 2013 (UTC)

(ec)There is not "the mathematical theorem", there is an infinite number of them (and a large number of known and interesting ones). If something is a theorem, you cannot refute it by definition - if you can refute it, it is at best a (failed) hypothesis. Typical ways of refuting a hypothesis are by counterexample, or by reductio ad absurdum. To refute e.g. Fermat's Last Theorem (and thus proving it misnamed ;-), you could just produce natural numbers $a,b,c,n$ with $a,b,c>0$ and $n>2$ such that $a^{n}+b^{n}=c^{n}$. --Stephan Schulz (talk) 11:52, 6 September 2013 (UTC)

(edit conflict) Finding a counterexample is a very good route. Finding an error in the proof is less good because although that proof may be incorrect, there may be a different and correct proof that has not yet been found. We can be more specific if you tell us which theorem (or conjecture) you hope to refute. Gandalf61 (talk) 11:57, 6 September 2013 (UTC)

However, finding an error in the known proof(s) can help persuade others that it is indeed the original theorem, rather than the proposed counterexample, that is incorrect.—Emil J. 13:20, 6 September 2013 (UTC)

True. So best route of all would be to find an error in the published proof and a counterexample that illustrates this error. Gandalf61 (talk) 13:42, 6 September 2013 (UTC)

Or if you cannot produce a specific counterexample, an existence proof that one must exist. -- The Anome (talk) 14:32, 6 September 2013 (UTC)

Where can I provide Mathematical proof  ?

Where can I provide Mathematical proof I want address and website37.238.55.76 (talk) 14:30, 6 September 2013 (UTC)

There is no single place where you can provide mathematical proofs. Most proofs are published at scientific conferences or in specialized mathematical journals. Typically, each journal will have its own way of how manuscripts are to be prepared and submitted. If submitted, they will be reviewed by one or more editors, and if they look promising, they will be handled to specialist peer reviewers, who will check them for correctness and significance. If the reviewers suggest acceptance, the journal will typically publish the paper, often after suggested changes have been made. Note that the standards expected of scientific papers are quite high - in my experience, it takes about a day per page to prepare them once the research itself is done (others may be more efficient than I am ;-). Also, it takes anywhere between a few weeks and a few years for the process to complete - in recent years, this has sped up, but a few months are to be expected (the Journal of Automated Reasoning typically gives me 56 days for reviewing). A simpler and more modern way is to publish a draft on arXiv. However, even there there are some barriers to entry. --Stephan Schulz (talk) 15:11, 6 September 2013 (UTC)