perfect play = draw??

http://en.wikipedia.org/wiki/Solved_game#Perfect_play

There's a few counterexamples to our resident troll's ridiculous post in there as well.

Yes, because the way they think are so radically different. Computer wins don't come from their opening understanding in the slightest. This isn't like humans of different styles, computers and humans could not think anymore differently. Ok, that statement is so false it's not even funny. It's clear that you're the one who doesn't understand chess! Lets see what everyone here thinks who knows more about chess.

Are you saying chess players should be taught how to calculate 20 moves ahead in every position? Because that's basically what you're implying when saying chess is a big tactic. In the extreme sense, yes if you can calculate 85 moves ahead it is, but that "plan" when playing another human as a human will fail miserably and you know it. Thus part of being a good chess player is knowing when and when not to calculate. If only humans could calculate like computers... a top GM would then be well over 3000.

+1.

No, I'm not saying that, because that's a very bad way for HUMANS to play the game. However, we're not talking about how HUMANS play, we're talking about PERFECT play, which potentially means calculating hundreds of moves into the future.

The human way of thinking about chess is so terrible that it loses 100% of the time against computers and their tactical way of thinking. So, what makes you think that your human way of thinking tells you anything about perfect play?

I'm not sure why people keep quoting this. Perfect play is well defined:

"A solved game is a game whose outcome can be mathematically predicted, regardless of the moves played. Games which haven't been mathematically solved in this sense are said to be "unsolved". Solving a game generally involves assuming both players will always make the best move possible (that is, the move which, in the long run, will ensure they either win or draw that game).

In game theory, perfect play is the behavior or strategy of a player which leads to the best possible outcome for that player regardless of the response by the opponent. Based on the rules of a game, every possible final position can be evaluated (as a win, loss or draw)."

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

It may not be absolute perfect play, but the point is it's close enough. The only thing we have to be accurate in is the assesment of key positions in certain openings. Now, in sharp positions, it might be tough, but do you really think quiet positions (I hope you know what that is, since you know more about chess than me) considered equal by gm analysis are over 1 whole point off???? Remember, it only takes one "sound" black opening to show that chess is a draw. In an actual game, yeah these GM's would lose most of the time but that's despite knowing more about positional play and thus the assesment of opening positions.

"Perfect play" in irrational, complex positions, we can't come close to, but for the opening positions we only have to be within 1 whole computer point (I mean a 1400 could probably avoid getting losing opening positions if the opponent plays quietly).

Yes, I cited that link myself earlier today when I said that "perfect play" was objectively defined.  The point isn't that it's not well defined, but rather that the term perfect play is continuously misused in these threads to the point where the dialogue has no bearing or relevance to actually solving chess or what the solution might be.

No, no it's not.  Proof is proof, evidence is not.

How is it not close enough? Name some lines where the GM's assesment is completely wrong or he's missing some move.

I'm sure there are many lines where GMs play very well, but have deviated from perfect play.  I can't name them, but more to the point, you can't name the lines where GMs have not.

 The problem here is that for all human players, even GMs, there is quite a difference between "open" and "closed" positions. Calculation is much easier (though certainly not easy!)in open positions because there are many lines that can be refuted with relatively simple variation trees, although it still takes a lot of deep analysis.

 But in closed positions there are too many possibilities that look promising, and GMs must instead use intuition or their knowledge of opening theory since the thousands of variations are too numerous to be effectively analyzed.

 However, if White were perfect, he could theoretically explore as many of the 10^120 or so chess positions as necessary to ensure he always made the best possible move from the very beginning. So if there was initially a mate in 270 for White, he would see it at the outset and would make all of the correct moves to force the mate in 270 moves! One winning line in so many trillions of possible games seems very likely to exist, and that's all it would take to prevent Black from drawing.

sorry but it seems you just don't know much about the game and you just don't know what you're talking about. Actually, I doubt anyone who finds it likely that white has a forced win knows much about the game.

wouldn't it be hilarious if after chess is solved that white has a forced win afer some silly moves like philidor_position suggested.  for all we know 1.a3 could be a forced win and no others are.  generally though it seems to be a drawish game.

Yeah, some of these people are just guessing without knowing about chess. They may be educated guesses, but you're likely to be wrong when you have such little chess experience, which cobra and the other guy obviously have. Lots of weird things happen in some chess positions, but those brillinat tactics have never been forced, it has always been due to a blunder or mistake that gave the person the weird tactic.

And cobra correct me if I'm wrong, but if you have been listening to what I've been saying about black's opening choices, clearly there isn't just one specific line white would have to prove, he would have to prove he is winning in every single opening, because it's obvious something like scholar's mate can't be forced. I don't think "one line" is the perfect word, even if it starts with only one opening move.

i just find it hard to believe that there is one move white can make where he wins no matter what combination of moves a 'perfect' black opponent subsequently plays

How about this. White has 20 possible first moves, which means that Black needs 20 perfect defenses in order to hold a draw. White only needs ONE perfect offense in order to win. Therefore, it's 20 times more likely that perfect play will lead to White winning rather than a draw.

Yeah, if all of white's moves were as forcing as e4 or d4. There is no question that the more passive choices (anything other than d4, e4, c4, and Nf3, but really e4 and d4 are the main dangerous options to consider a forced win with) will not win by force. But of course, you assume they are all equal. If you figure playing 1 a3 is virtually playing black for example, you can eliminate that right away. Black needs just one possible defense for the most dangerous options, any other first white move is so obviously not giving him the initiative (or at least it SHOULD be obvious) that it doesn't even have to be considered. Tactics just don't come out of nowhere in black-like setups. White's one perfect offense has to beat all of black's defences, so you can split that one offense into many. Like I said before, if there isn't some crazy scholar's mate working by force, that's what will be needed. There are ALOT of defences to e4 as well as d4, the only ones close to being losing would be the KID (maybe some ultra deep line, though still HIGHLY unlikely) and maybe some sicilian line (but it's even still very unlikely). Our knowledge of endgames has shown us that you need to be able to get certain advantages in the endgame to win, so no small edge will do.

Due to heavy computer analysis, it seems that chess is a draw with perfect play.  This has not been proven yet however, but the probability it is true is extremely high.

P.S.

Tic-tac-toe is a draw.  In Connect Four, the first player wins.  Checkers is a draw.