True, but as you say most people read the first few pages and stop reading. If what people read are the first few pages, why wouldn't you want a filtering system to determine highly rated posts.
How do you KNOW they are "highly rated" when you have NO IDEA if they are ALSO highly hated or down rated or whatever? My point is that you could have a controversial issue where 100 people "like" a post and 5000 dislike it. But with this system, you'll NEVER SEE the dislikes and THAT is why it's a TERRIBLE SYSTEM. It distorts the truth by not representing both sides of the issue.
Posts that are disliked, do not get upvoted. This naturally filters out the bad posts (the reason why you don't need a Dislike button in the first place).
No, it doesn't because a bad post as indicated above would have down votes that could EASILY OUTNUMBER the up votes by 100x even. How can you possibly get an accurate idea of the "quality" of a post by a system that ignores 50% of the voting issue?
And NO, a non-vote is NOT even REMOTELY the same as a "down" vote. I don't up-vote every post I think is "good". I only up-vote ones I strongly AGREE WITH (not the same thing even remotely as a "good post" as it's not based on how well written the post is, but whether I want to endorse a VIEWPOINT. Even then, I don't BOTHER do vote for 99.9% of the posts as it's too much work to vote on a 5000 post thread and THAT is why I'm saying a few initial votes appear at the start and then they disappear as time goes on. Yes, some people only read to page 3 or whatever, but clearly a 200 page thread has a lot of people still reading and posting to it. But you won't see more than 1-3 votes on such threads for the most part. This is CONSISTENT over time. And that comes back to the "don't want to bother" aspect once again.
When we had up and down votes there was still a problem and that was it only showed the total sum of the posts (i.e. 2000 down votes and 2001 up votes would show up as "1 up vote") and that didn't say a lot either. What it needed to do and what it should do is show the total up and total down votes separately and that would tell people how many people liked and disliked the post enough to bother to vote. No votes doesn't indicate a bad post. It indicates a lack of strong agreement or disagreement. But even if I disagree with a post, that doesn't mean it's bad grammar or poorly written. Most people "like" posts they agree with. Such posts could be one sentence. Your assumption it's a "quality post" is simply not supportable from what I've seen. People agree with the first couple of posts that represent their position on a given issue/review/model/whatever and then ignore the rest except where exceptional points are made and even then it wanes over time.
What one should look at is why the "down" votes were removed in the first place despite polling indicating that a majority either wanted the downs to stay or didn't mind them either way. A small extremely vocal minority threatened to remove their voluntary monetary funding of the site unless the down vote was removed (probably because most of their posts were getting down voted all the time). I personally despise political correctness and baseball games where the score isn't kept so everyone can "feel good". It's a load of horse manure as it doesn't represent reality and on a larger scale it doesn't prepare young adults to deal with that real world (where terrorists don't care if you feel bad when they light you on fire in a cage).
A post that's been upvoted many times is liked. Whether it be useful or simply funny, it was enjoyed. I firmly believe that's important.
Like I said, it's only liked by those that up-voted it (and bothered to do it). How many disliked it? If the original poll to get rid of the system had only showed votes to get rid of it then we would have seen what appeared to be 100% of the voters opting to get rid of it and hundreds of them at that. But does it matter that 50% more voted to keep it in reality? Do you see what I mean by a total distortion? Your assumption it represents "liked" posts is simply wrong. 10 likes and 1000 dislikes is NOT a liked post. Wipe out the 1000 dislikes and you get the appearance of overall people liking it where that is simply not the case or you can assume with 200k people on this forum that only 10 out of 200k "liked" it but that would be a completely inaccurate assumption also as nowhere near that amount will read that thread and most people won't participate either way (either not caring or not bothering and thus the only likes you get are from a tiny minority that enjoys "liking" posts).
I'd prefer no "like" system over a terrible system and this is a terrible system, IMO and now your name is attached to it as well. Imagine if anyone could look up how you voted on an issue during an election. Your neighbor might suddenly hate you when he finds out you voted for/against gun control or abortion or whatever when it's none of his business. Thus, I find the system more flawed than ever as it adds that aspect to the above mess as well.