Almost four years after this discussion, and I think that the current situation is not ideal and we need some rethinking. Users keep tagging with questions that should be tagged with just .

Would it be feasible, eg, to add some warning when people with low reputation try to use this tag?

• A warning can be added, but not based on reputation. SO has a helpful tooltip for sql tag, for example. Also, my impression is that this attempt to differentiate probability questions by level has largely failed and can be abandoned by merging the tags. – user147263 Dec 4 '14 at 15:17
• The idea of separating like like calculus/real-analysis was nice, but the tags' names are too close for this to work IMO. It's not intuitive at all, and people hardly read the tag excerpts -- especially when they are this long. – Najib Idrissi Dec 4 '14 at 15:40
• I suggest reformulating the tag excerpt to emphasize "Do not use this tag for questions about how to compute a probability. Instead use probability in this case." Users who mistag can then be specifically pointed at the excerpt and the tagging can be fixed. It seems a little intrusive to create a pop-up, specifically looking at the list Maybe a context-sensitive pop-up should appear if multiple probabilty-* tags are added. It seems to me that most mistagged questions are tagged with many (or all) probability-* tags. – AlexR Dec 4 '14 at 16:38
• Side note: Most mistags also seem to fall to users with <1000 rep, so that looks like a good threshold to check for warning if that is possible. – AlexR Dec 4 '14 at 16:40
• @AlexR Good point about the overlap. There are $7719$ questions in (probability-theory), of them $3131$ are also tagged (probability). That's $40\%$. I think when a tag is used incorrectly $40\%$ of the time, it's about as useful as not having it at all. – user147263 Dec 4 '14 at 16:44
• I'll just point out that there is also a comment template which can be used to notify users that they are using the tag incorrectly. – Martin Sleziak Jan 1 '15 at 8:13

I restate my point of view as an answer, so that it can be voted both up and down, unlike comments.

There are $7985$ questions in , of them $3221$ are also tagged (probability). That's $40\%$. When a tag is used incorrectly (at least) $40\%$ of the time, it's about as useful as not having it at all.

Therefore, we should recognize that the attempt to differentiate probability questions by means of two tags and has failed, and merge the tags.

• I think merging is too drastic a measure. Renaming the tags to separate them more clearly would have my preference. – Lord_Farin Jan 1 '15 at 11:24
• @Lord_Farin What would these names be? It doesn't look like adding "elementary-" prefix worked out very well for set theory or number theory -- questions have to be retagged very often. There are 1732 questions tagged number-theory and elementary-number-theory (but only four tagged set-theory and elementary-set-theory) – Najib Idrissi Jan 1 '15 at 12:12
• @NajibIdrissi I have to think about that; when I find some suitable candidates, I will post an answer. – Lord_Farin Jan 1 '15 at 12:13
• @Najib: There was a substantial overlap between the set theory tags, but the issue was somewhat raised by Arthur Fischer and with a combined effort of several users we managed to reduce this and keep the intersection to a minimum, without flooding the front page too much. – Asaf Karagila Jan 1 '15 at 13:23

It is manifest that there is a problem. I think that part of this problem is the names of the tags -- they are just too similar. My suggestion would be to rename them; my proposal:

I came to this suggestion by considering what words really separate these categories, and concluded the first is always supposed to be about computation (be that numeric or algebraic), while the second is supposed to be about the theoretical part of probability theory.

An added advantage of this convention is that the theoretical tag doesn't start with "probability" anymore, which will hopefully reduce the number of users thoughtlessly adding this tag to a question. Plus, I conjecture that someone not really comfortable with the subject will likely avoid a tag containing the word "theoretical".

Naturally, I warmly welcome other name suggestions.

• Tags are sorted by number of questions when you input them on a question, not whether they start with the word or not; try to make a new question with "set" in the tag box and see what I mean: elementary-set-theory comes first. And new users have the tendency to put in as many tags as they can on their questions. – Najib Idrissi Jan 2 '15 at 9:24
• I was thinking that probability should become elementary-probability, and probability-theory could either be left as-is, or perhaps something like probability-analysis. I do think that questions regarding probability spaces, measure-theory, etc. don't belong in the same boat as "pick k balls from an urn..." – Emily Jan 2 '15 at 20:00
• @Arkamis: You are probably right. (Sorry I just had to make that pun!) – Asaf Karagila Jan 5 '15 at 0:19
• @AsafKaragila Almost surely. – Emily Jan 5 '15 at 3:31