How do you search for duplicates?

Question

Sometimes I am really surprised how fast are the duplicates of new question found, given that we have vast amount of questions here.

I hope posting this question here can be useful and can lead to some constructive suggestions. If it was discussed here before and I've overseen it, my apologies. (It would be, however, a little ironic.)

A few methods for finding duplicates that come to mind are the following:

• Look in the list of related questions, which is generated automatically.
• Sometimes it is useful to look in this list: http://meta.math.stackexchange.com/questions/1868/list-of-generalizations-of-common-questions Another useful list is mentioned in this post: http://meta.math.stackexchange.com/questions/16708/catalog-of-standard-exercises
• If you remember that you posted such question/commented such question/answered such question, it might be relatively easy to find it in your list of comments/answers/questions.
• Even if you only remember seeing such a question, it might make your life easier - you can remember poster, part of the title, a phrase used in the answer, simply something which helps you find the question more quickly.
• You might search for the question - either here or using google. This might be not that easy if the search would involve mathematical formula, but:
  
  • If the formula has a name, it is probably mentioned in the answer or in the comments, so you might try to search for the name, e.g. "triangular number" site:math.stackexchange.com on google or triangular number on MSE. (The search on google has the advantage that it also looks into comments.)
  • If you know the correct result, you can search for some part of formula e.g "frac n(n+1)2" or "n(n+1) 2" site:math.stackexchange.com on google or here. However, the search for "n(n-1)/2", "k(k+1)/2" or "s(s+1)/2" might be reasonable in this case, too. (In general, it is not easy to search for formulas.)
• If it is a question which was probably asked many times on Math.SE, you can try your luck on the frequent tab, which shows the questions with most links. (Perhaps it is useful to restrict this to some particular tag, like here.)

Are there some other useful tricks to find the duplicates?

Answer 1

Here are some techniques that I find very useful. When composing posts I explicitly insert keywords that will aid me in quickly locating the post by searches. For example, I strive to include commonly attributed names of theorems or identities, and when such names have alternate spellings I include all of them, e.g. l'Hôpital and l'Hospital, or Besicovitch and Besicovic, etc. This has the added benefit that I can quickly compose links to my prior posts on related topics by using searches employing these keywords, e.g. the telescopy buzzword.

If we all make some conscious effort to ensure that posts contain useful keywords then that should go a long way towards making searches more effective - whether they be for locating duplicates, discovering related results, etc. Of course the built-in SE tagging system should be the first level of categorization used. But such free-text tagging by keywords proves very useful in situations where the SE tagging model doesn't work well, e.g. in answers that mention topics that go beyond that categories in the original question. No doubt others have evolved analogous useful techniques. It would be very helpful to collect such community wisdom in this thread.

Answer 2

In case someone stumble upon this question these days, there is a great tool for finding duplicates since 2016 : Approach0. It's quite remarkable, it can for example find same expression with different letters being used in it.

For more information you can see this question where it was announced by the author: Announcing a third-party search engine for Math StackExchange.

My wish is that more people know about this before asking or even answering questions.