After reading a question and providing an answer on math stack exchange, I often find myself thinking about it from time to time and making more progress on a fuller understanding of the issues/problems/concepts or a better way to explain the answer/calculation. At what point is it considered overboard to continually edit answers to improve them?
There are no official restrictions. But a few comments/observations/pieces of advice:
- Frequent edits begin to annoy others at some point. Edits "bump" the thread to the front page of active questions, so if you give the impression that you are trying to hog the attention of others, then expect negative reactions. An old rule was that after edit #10 the post becomes Community Wiki - meaning that you will no longer get any rep points from it. This rule has been abolished, but at that point the system will raise an automatic flag. This will cause a moderator to investigate, and give you an opinion/advice. Or, in a rare case do something more drastic.
- (Questions only) Do not alter the gist of the question (or replace it with another one). This will annoy others (what robjohn says under Woodface's answer). The exception is that you should, of course, fix any clear typos and/or add the omitted assumptions and such. Particularly when the question is a fresh one.
- If you really want to fine tune your post a lot (improve LaTeXing, gradually add a lot of detail,...) it is a good idea to use the sandbox. It was created for this purpose. Read the instructions there (immersed in the question body).
It's not a problem unless your repeated edits begin to annoy people.
When they begin to annoy people, they will tell you to knock it off, using some combination of comments and downvotes.