This question already has an answer here:
- How does accepting an answer work? 1 answer
Something is wrong with answer ordering: different versions of Chrome (under different environnments) show different behaviour.
Noticed here: How to debug IE9 HTA?
- WinXP(VM) and Win7, Chrome Version 33.0.1750.154 m: constant order
- Linux, Chrome Version 23.0.1271.95: random order
Might extends to other browsers?
In the end it looks like a browser specific issue, in the very particular case of:
- a self-accepted answer with N votes,
- other non-accepted answers which with N votes too,
- sorting answers by vote,
- the browser is an old version Chrome, Firefox (which version?), maybe others too.
In this case, the order of the answers with N votes varies randomly at each refresh, including the accepted one. This doesn't happen on more recent versions of Chrome, accepted answer is always first and the order remains the same between refreshes.
I understand the choice of using random for ordering equally voted answers: a simple way of giving the same visibility without the hassle of taking rep or time into account, etc.
However I had the case where I needed to refer to an accepted answer which had the same number of votes than other, less acurate answers, and was surprised to have to search it each time I reopened the page (there can be many answers with 1 vote).
See this question for example: How to debug IE9 HTA?, where the accepted answer is not ordered at the beginning.
Do accepted answers deserve more visibility than others when votes are equal?