After spending quite a bit of time trying to crack a problem about the convexity of a multivariate function by myself, I asked this question a couple of days ago: Convexity of difference of log-sum-exp: $f(x_1, x_2, x_3, x_4) = \log(e^{x_1} + e^{x_2}) - \log(e^{x_1} + e^{x_2} + e^{x_3} + e^{x_4})$.
A user (one of the top users in the math.SX community) did a great job helping me find the answer. I actually learned a useful lesson: when trying to analyze convexity of a multivariate function, start by restricting yourself to a subspace by fixing some parameters.
This user, however, instead of writing an answer, solved the problem in a comment. When I suggested him to transform his comment in an answer, he suggested me to delete it (and it correlated with a downvote on my question, although I cannot know for sure it was him / her).
I'm fairly new to the math.stackexchange community, so perhaps I misunderstood the kind of questions that are expected here. I read Differences between mathoverflow and math.stackexchange., which states that math.SX is a
Q&A site for people studying math at any level & professionals in related fields
My question is definitely not research-level, and turned out to be fairly simple to solve (it always seems obvious once you have the answer...)
Even though it might seem very ad-hoc at first sight, the question is motivated by the general form of the likelihoods of some statistical models of comparisons / choices, such as elimination by aspects and paired team comparisons. (Note that I did not talk about this connection for now in the question - this could certainly improve it).
So my questions are:
- is the type of question I asked not welcome on math.SX?
- if not, why exactly?
The comment "because it's of no use to others" seems a bit subjective. I certainly know other people who would be interested in this question & answer. Perhaps "not enough others"? But what is enough?