perfect play = draw??

I don't think you can resist comparing chess to all that other stuff, but just because they seem similar doesn't mean you can compare them. In chess in every opening (I can't emphasize it enough) that GM's don't think black is lost there has to be some move that none of them saw that somehow gives him a losing position. And like I said, that has absolutely never even came close to happening with "perfect play". So you need a couple 100 of refutations totally overlooked, and that goes for the quiet positions! Those quiet position refutations will be quite hard to find!

We do? If these grandmasters are so great, why are they so easily trounced by computers? Clearly, grandmasters have no idea what the best moves are and play the wrong moves all the time.

Grandmasters dominate amateurs, computers dominate grandmasters, is it not reasonable to assume that perfect play would dominate computers? Why do you believe that current theory is "at the top of the food chain", so to speak? Theory is constantly changing from one year to the next. Everything we knew about chess 100 years ago has completely changed. It will probably change in the next 100 years.

How can you possibly say that your opinion on perfect play being a draw is anything short of a wild guess?

You know, I don't actually disagree with you. I think chess is a draw. It's obviously extremely likely that it is.

However, to claim something about perfect play, you need mathematical proof.

And for that it doesn't matter how much evidence of this sort you pile up -- all it can ever do is make it more likely that it's a draw, but it can never make that final step from "extremely bloody likely" to "certain".

I disagree with those who have said that the outcome of perfect play is a draw. It must be kept in mind that if White having the first move gives him ANY kind of advantage, or disadvantage, then this margin (no matter how small or subtle it is) of inequality can, with perfect play, be exploited and gradually widened, even if the other side plays perfectly as well and the win requires well over a hundred moves. If this slight advantage is ever relinquished, then by definition the "perfect" player must have made an error. So theoretically, chess is only a draw with perfect play if the inital position is ABSOLUTELY, positively equal (+0.00000000000000000000000000000000000001 at the outset does NOT count as equal), which I find very, very unlikely.

But by the same logic you use to say that if the advantage is relenquished then the White player must have made a mistake couldn't you also argue that if the advantage is widened enough to lose the game then the Black player must have also made a mistake?

 If, for example, White is left with B+K at the end and Black has just a K, it is not +3.0, with White "ahead" by three pawns but unable to win by checkmate. The position is simply equal and the game will inevitably be a draw. If neither White nor Black can win with perfect play then there is no advantage. Likewise, if there IS an advantage of any kind, it can be converted to a win by force by playing perfectly.

But what makes you think that white's advantage isn't ALREADY enough to win the game from the very first move?

Let's assume that white really does have an advantage (although this in no way can be proven). One side is arguing that the advantage is too small to win, and the other side is arguing that the advantage is enough to win. How the hell can any of you purport to know? Games have been won and lost because of a single pawn, sometimes even less like a simple positional consideration. To think that you are able to divine the outcome of such a slim advantage is preposterous.

It's entirely possible that white's advantage will lead him to being up two knights at the very end, and yet unfortunately result in a draw. It's also possible that white's advantage will lead him to an astonishing queen sacrifice. Who knows.

Uh! This is why it infuriates me that you debate about these things without actually playing chess so much!!

Ok, computers do better than GM's because they can calculate millions of positions per second. Instead of considering the advantages and stuff, they just calculate as much as they can the best line they can find many moves deep. They lack the plan and positional understanding GM's have. However, their calculation usually comes out on top because it never tactically blunders, is extremely punishing, and sometimes the calculation finds out the winning plan (it sees that it wins a pawn or something). However, although they can beat GM's, that does not mean they know more about openings than we do. GM's understand the position, and know why one is better or worse. Computers despite their lack of understanding (which is why its assesment is not in the books unless it detects a blunder) will still come out on top in a game because of the practical convenience of being able to calculate so well and easily. Yeah, try being taught by a computer, man. And why don't you actually play alot chess for several months before debating this stuff?

"If, for example, White is left with B+K at the end and Black has just a K, it is not +3.0, with White "ahead" by three pawns but unable to win by checkmate."

There are exceptions to every rule. However, if one is up a piece in the typical manner, by winning it out of the opening, and the opponent has no compensation, then that will not happen against correct play (perfect isn't even necessary). A piece can eventually put enough pressure on the opponent to win a pawn, and that pawn can easily be guided with an extra piece or at least force the opponent to sac a piece for it.

I've already stated clearly that I don't "purport to know".  I'm afraid your outrage is misdirected.

I suspect, given the sizeable imbalances in the endgame that still result in draws, that White's first move advantage is unlikely to be sufficient to leverage into a forced win.

It is well known by actual CHESS PLAYERS that it takes over +1 to force a win or be winning. If the advantage at any given point is less than that, then indeed, in 70 correct moves later or so it will, eventually, dissipate into 0.00 (meaning draw, not necessarily equal, but that a draw will happen for certain, this only really applies to endgames. Think draw by insufficent material).

And Scarblac, everything I have said may not be extremely accurate, but my point is for people who have actually played chess and are good it's obvious to them how and why chess is a draw. And then there's that .01% chance or so that we're wrong.

AHHHH 2012 THE WURLD IS ENDING.

D:

So, basically what you're saying is that the poorer player knows more about the game than the better player? Following your logic, based on our ratings, I know way more about chess than you do.

You don't seem to understand that "strategy" and "understanding" are simply crutches that we human beings have to fall back on because we don't have the computing power that computers do. However these crutches are not GOOD things, they are in fact BAD things. Chess is nothing more than one big tactic. That's why computers are crushing, and that's why we cannot have any glimmer of insight into what perfect play will lead to. Perfect play has nothing to do with strategy.

How would we know if the game was perfect

After 12 moves there are more possible continuation than known stars in the Universe.

a Super Computer would run forever tying to solve Chess.

It is well beyond the human Mind.

A regular grandmaster today absolutely knows way much more than the best engine (namely, Rybka 3), especially on openings. Rybka does have a pretty good knowledge base, but its main strength is that it just doesn't make any blunders.

How do you figure? Elubas' rating is five hundred points above yours.

I was at a friend's house one time and decided to test out his Shredder computer chess game. What i wanted to know was: what would happen if the computer was to play against itself?

The results shocked me: black won!! I decided to run the test again. This time, white won!!

Although Shredder is obviously nowhere near 'perfect play', the fact remains that chess is a very complicated game, and it is difficult to know 100% whether a perfect game will result in a draw, although it seems likely.

I simply don't have the time to read through ALL of the posts in this article and set those STRAIGHT who need setting straight. So, in response to the original question. Perfect play ALWAYS equals a draw, no matter what you are talking about.

Checkers has been solved by CHINOOK, resulting in a perfect game. In CHINOOK vs. CHINOOK games, all the games are a tie. 100%. That means perfect play.

ANALOGY TIME. Let's assume that you are a pro golfer with the uncanny ability to hit a hole in one on EVERY HOLE. Par 3,4,5. After playing 18 holes, your score would be 18. PERFECT. Unbeatable. If another person came along with your uncanny abilities and you played him or her, what do you think would happen?

Well, you'd both always tie. Every game would be 18 to 18. 100% time. This analogy is FOOLPROOF.

When chess does get solved within the next few decades, perfect chess play will become routine by both men, women, children, and MACHINE.

Yes -- I've already memorized the six-peice tablebases and those portions of the seven-peice tablebases that have been completed.  The 32 peice version should be a cake-walk.

what is your definition of a perfect game?

It's not subjective -- perfect play is objectively defined as those moves that lead to the best possible outcome no matter what the response is from your opponent.