It is a not unheard of phenomenon that answers don't get much upvotes in the following scenario:
First, a user asks a non-trivial question, which turns out to be non-trivial enough not to get an answer for a while.
Then, the question lands in the big heap of unanswered questions, where it does not receive much attention.
Finally, some user digs up the problem, devotes some of his time and energy to come up with the answer, and posts it. By now, it's probably only the OP who pays any attention (or maybe not even him).
The answer gets accepted, upvoted by the OP, and forgotten.
The net result is that much work is rewarded with some +20 rep, or maybe less. A situation in this spirit was discussed for instance here (but in a somewhat different context). This is in contrast with what happens with comparably easy problems where an immediate answer can earn more rep with much less effort.
I do realise that upvotes are not precisely the point of MSE, which is to generate good quality answers to mathematical questions. However, I am afraid that some users may be discouraged from posting late answers, because answering the easy ones "pays" better. This could result in there being a lot of answers to the easy problems that don't bring much new to the site, and with some interesting problems being left unanswered.
Is there some mechanism that could help such situations to occur less freqently? For instance, if there was an accessible list of questions where the question got much more upvotes than the accepted answer (perhaps with the extra condition that the answer was given relatively late), I would be happy to look through it and upvote under-appreciated answers if I happen to see any. But such query is rather too specific for the MSE search engine, isn't it?
I see that the "review" tool has the category "Late answers", but it apparently applies only to the ones by "new users", and seems not to be used much. (I, for one, have no possible reviews in this category.) Perhaps it could be considered to extend this feature? For instance, "Late answers" could also include late answers from experienced users, answers that got relatively little rep (my gut feeling is that if question deserves $x$ upvotes, then an accepted answer generically deserves at least $x/2$) or little attention, and so on.
(Disclaimer: I do not hide that I posted some late answers recently. It's not that I feel they are of particularly high quality or that they are not appreciated enough. It's rather that this served to direct my attention to the general possible problem.)