Firstly, these are mechanisms that are deeply embedded in the StackExchange platform, so there is no way for us to handle it differently just for one site.
This leads us to Meta.StackExchange - and we can see that this idea has a long history there (you should read the comments to the answers, including follow-ups, as well).
In short, this has been discussed at length with community managers involved and apparently lead to nowhere as of yet (see tag status-declined).
Secondly, how answers are received should really not be any standard for evaluating the quality of the answer in terms of being a good fit for StackExchange. The resulting summarised votes in effect should, but due to the relatively small basis of expert members (both in terms of understanding the purpose, rules, and mechanics of StackExchange and in terms of a knowledge base in philosophy), they not necessarily do on this site.
If you answer to a question that is a bad (or even a good) fit with an answer that is a bad fit (e.g. advice, opinion, obscure position) and helps the questioner personally, does this make the answer a good answer that should be upvoted on StackExchange?
No, definitely not. StackExchange aims for a database of objective knowledge, not a collection of personalised and individually addressed dialogues - there are in fact other platforms for that.
Thirdly, to the specific case that triggered your request. I tried to read very carefully (took three different statistical undergraduate courses) and in my opinion, it really is not that good an answer.
It dips into statistics but leaves out the highly related number theory behind it: On one hand, "exact" numbers can in fact be intervals e.g. the example "0.31415" is in fact the interval between (including) 0.314145 and (excluding) 0.3141549¯ (periodical 9), allowing for probabilities. On the other hand, periodical numbers (which in a very particular mathematical sense are "exact") then become a different problem - how to measure that we really observe it and not an indiscriminately smaller or bigger number? The answer addresses the first aspect in the solution to the puzzle, but only implicitly.
Also, the mathematical descriptions are much more extensive than the addressing of the "philosophical" core of the problem that could be discussed (How do we get these probability functions? What are their reliability and validity - is it ever 100%? Why can a theory never supersede empirical factuality? Can exact values become empirical content at all?).
There are a number of authors that address these very questions explicitly and could have served to construe a referenced, well-founded answer. The theory of counterfactuals by Lewis and several authors in the philosophy of science come to mind. But instead, we have a narrative that kind of misses the point - philosophically, that is. Therefore, it is not really a good fit for Philosophy.SE. Thus, I do not even see any good reason for retracting the specific downvote despite the edit (disclaimer: was not mine).