Should I upvote an answer based on how much it was improved?

Question

Should I hold how much an answer's score increased/decreased (depending on the winning criterion) as an indicator of how good it is and whether or not I should upvote it?

I used to see answers golfed by a hundred bytes as better than others even if they were longer because it seemed they were worked harder on and probably had reached the limit, but I realized maybe some people golf their code as much as possible before posting and maybe worked even harder, but I wouldn't be able to tell. Do a lot of people do the second, and if so, should I also upvote answers without strikethroughs that are already as short as their language allows?

Answer 1

No one can tell you how to vote

Voting is a very subjective thing. Some people are liberal, others not so much. Ultimately, you should upvote or downvote an answer (or question) if you think it's deserving of that vote.

This faq post on Mother Meta is a very similar question, asking when should you vote. The accepted answer says

I generally use "does the post help the site?" to judge upvotes and downvotes to some extent.

which I personally think is a good criteria for voting. Golfed answers (or more generally, answers that have improved their score) certainly help the site, after all, that's what we're all here for. On the flip side however, there are certainly users that golf before they post, not after (myself often included), and so only voting up answers with demonstrated golfing may not be the best idea.

Ultimately, if you think it deserves an upvote, give it an upvote. If you think it should be downvoted, downvote. If neither, do neither.

_{Note: people will be constantly improving their submissions from the moment they get a working program to the point where they (and others) sit back and believe it's an optimal solution. Sometimes you'll see these improvements, sometimes you won't, which is why (IMO) you shouldn't vote just because an answer has struck-through scores. If, however, you want to, no one's going to stop you}

Answer 2

I think the direct answer ought to be 'NO: you should not upvote an answer based on whether it was improved.'

If this was common practice, it could encourage answerers to deliberately submit suboptimal answers, just so that they could gain upvotes by subsequently 'improving' them. This is (I think) certainly not useful nor desirable.

I must admit that it had never crossed my mind that an edited/improved answer could be considered 'better'. Actually, although I frequently do edit my own answers when I think of a way to improve them, I generally feel rather sheepish about it inside, and usually regret that I hadn't thought of it before posting them in the first place...