Wat. Seriously, any of the top 10 players would be top correspondence chess players pretty much immediately if they so chose. There are very, very few skills unique to correspondence chess when compared with OTB chess. Just because postmortem analysis shows that top players will sometimes make significant mistakes OTB doesn't mean they're poor analysts.

Oh, and lol at running 400m and 5000m being remotely analogous. Fast- and slow-twitch muscle fibers ldo.

Haven't you read up about fast and slow brain matter doc

Any of the top 100 players in the world would easily be champions. Their level of pure chess understanding is so much greater than the best correspondence players that it's absurd. I know guys whom have been the top correspondence players in the world, it's a complete joke to even put them in the same sentence as a top flight GM, regardless of what type of chess you are playing. (and they would all tell you the same thing of course)

The top correspondence player in the US is rated freaking 1800 USCF!! Number 2 is a bit better at 2242, but cmon. Maybe they could beat me (although I'm sure that if I worked hard I would be better than them eventually), but to fantasize that they could beat some of the greatest chess geniuses in the world is so insane. Top players don't play correspondence chess because there is no prize money, fanfare, or really reason to play. It's like talking about how some guy at your local 1-2 holdem game could beat the 1000-2000 at the bellagio. Let me tell you if some correspondence organization puts up a 100 million dollar prize fund instead of like 3k for first for a tournament that takes 10 years to finish, I'm pretty sure the names at the top will change.

curtains,

The bullet champion on ICC a few years back was a player from my club. He was rated around 1900 USCF. Now, the bullet championship on ICC is something that a lot of very strong players are into. Why then didn't a top GM win? Could it be that it's stupid to use someone's OTB rating to predict how well they'd do at a radically different time control? Hmmm...

This may have stretched the cc players are lol bad a bit far, even though I completely agree with your general point. Ulf Andersson is one of the top (2nd?) ICCF players in the world, although I'm not sure exactly how active he is. He was top 5 or 10 in the world OTB not that long ago.

Ok, when the player is an actual real GM or even strong IM then I admit they are probably pretty good at correspondence and would be expected to beat someone rated maybe 150-200 elo points stronger than them just due to having experience with the game and all. (although would still lose for sure against someone like Anand, Kasparov.)

1 0 on ICC has so much less to do with actual chess and more with moving the mouse fast (although still on average the higher rated players are much better). Next you are going to tell me that the best 5 second player in the world is rated 1600.

Ulf Andersson is supposedly 1 or 2 in the world right now in correspondence chess. Surprisingly when he was an actual active regular chessplayer, he didn't play tons of correspondence chess. Now that he is no longer anywhere near the top of the world, and is getting older, he decided to try that and lo and behold he is one of the best in the world, which actually proves my point pretty well.

It's an interesting question, but eventually chess engines will win. They will eventually play perfect chess and will always either draw or win - a human will not be able to be perfect.

Endgame databases will be perfected to the point that a marginal advantage in the middlegame will yield a win for the computer (it's already the case but will only be amplified as technology improves.)

Whether that will happen in 5 years is the only pressing question..

fwiw: I'm a Class A ranked player (USCF)

No, they won't "eventually play perfect chess" unless we somehow perfect quantum computation which I personally don't find likely. If computers ever play "perfect chess" it will be 100+ years from now. It's not like all it would take is a computer that's 1,000,000,000 times faster than the current fastest super computers. Such a computer still wouldn't play anything close to "perfect" chess. Don't confuse chess with checkers. Computers may be practically unbeatable in 15 years, but their chess won't be perfect.

Yes, a perfect playing computer is not coming any time soon. The number of games of chess (of up to 40 moves) is estimated at around 10^120 with around 10^50 unique positions. The number of atoms in the universe is somewhere around 10^80. Brute force solving chess is probably many centuries/millenia away. It's certainly not in the even remotely foreseeable future.

Your assumption is based on current technology; both hardware and software. No where in my post did I say that it will happen in 5 years, I said that it will eventually happen - which it will. You (and others with the same opinion) are assuming you can effectively predict the future which is laughable at best. You're quantifying the unquantifiable.

It will happen. Much bigger feats have been accomplish when all the critics scoffed at the perceived "impossibility" of said feat. Technological advancement isn't a linear function, especially with respect to the computer.

/my .02

In terms of resources required to solve the game, saying that chess is to checkers as checkers is to tic-tac-toe is an incomprehensibly large exaggeration in favor of checkers. (One could say the same about go and chess.)

You have no idea just how big a feat you are talking about. And yes, it is quantifiable.

OK, damagor, how fast will computers be in 25 years?

This keeps getting better.

Suppose they keep doubling at Moore's Law rates, which they won't because they already aren't. The computing power of the world still wouldn't be close. The relevant arithmetic is left as an exercise to the reader.

I suppose if we assume the only thing that will change in the next 25 years is processing power, you might have a point.

http://en.wikipedia.org/wiki/Paradigm_shift

Dude. Look at the numbers. This isn't some random "LOL, bots will never be better than humans at online poker type comment." The storage space alone for the table would be pretty much literally out of the universe. Assume you can store 1 chess "game"/sequence per "atom", which is already a beyond ridiculous statement. The physical size of the storage device would be larger than the entire universe. This isn't even starting to consider the time/power required to calculate that. And you can go even smaller. Even 1 position per quark and you still wouldn't have even remotely enough space in all of the universe.

I mean I think you don't really understand the scale we're talking about here. If computers become a trillion times faster than they are right now with a trillion times more space, nothing would change at all.

Dire, I think it is you that don't really understand what I'm getting at: Computer won't always be purely brute force..

http://en.wikipedia.org/wiki/Paradigm_shift

Then I think you don't really understand what playing perfectly is. Without brute force, it is impossible to play perfectly.

It would be possible for computers to play perfectly, just not possible to prove that they play perfectly. I'll leave this as an exercise for the reader.

Ahhhhhhhhhh, this particular error tilts me so hard. You meant:

This isn't some random "LOL, bots will never be better than humans at online poker" type comment.

Okay, resume the discussion about whether and when computers will solve chess.

oh ffs, that's nitty. I know where to put the qu"otes" but I'm too lazy to double check.

Possible, but I think beyond improbable. As endgame table basis show, perfect play is frequently very far from the natural course. And if it's not the natural course it can't be expected to be heuristically determined with a decent frequency, and if it can't be heuristically determined with a decent frequency then you're left with brute force. And if you're using brute force, but without analyzing all variations to their end, then you get horizon effected and are probably worse off than when you started just trying to use heuristics.

what if humans evolve into computers?

http://en.wikipedia.org/wiki/Buzzword imo