Solving chess

Solving chess . . . yeah right!

0.15 is NOT enough of an advantage to force a win.

Even a 1 pawn advantage can lead to an endgame draw.

And you underestimate the complexities involved in atomistic simulations at the macroscopic scale.

Anyway, let me try another approach. Matthew11, if the computer knows with 100% certainty that the beginning advantage for white is +0.15, and if your theory is correct and this means that white will win with perfect play, then if you were to play two computers with the same engine against each other, you should always get a white victory. Right? But do you see this happening? No, you don't.

The reason for this is that the computer does not know. A computer can only tell us what we program it to tell us. Until a computer can show a forced mate in # on move 1, all evaluations are based on position, not on checkmate. Whatever value a certain position gets is based on an algorithm programmed by humans. The computer isn't smart enough to make a direct link between positional value and checkmate. In other words, a computer can only win when it finds a forced mate; otherwise, it must rely on our kindness in providing (programming) the most helpful (accurate) way to assess chess positions, which can eventually lead it to finding this forced mate.

Thus, saying that an advantage of +0.15 dictated by the computer is game-winning for white would be equivalent to saying that humans have solved chess. This is clearly not the case.

"1. We don't know that this supposed 0.15 advantage exists. It is a GUESS made by the computer, because it cannot calculate the true advantage. Therefore, we don't know that white has any advantage at all."

Why do you keep refering to "0.15"? I'm more talking about the 1/3 pawn. Anyway, you seen not to be listening so I'll say it again. White has some advantage. If it's 0.15, 1/3 pawn or 1/2 pawn, it's there. With white being perfect, there is no way to take away this advantage.

"2.If white DOES have an advantage, then by definition, this advantage must be one which is enough to win with, and therefore equals infinity, or "mate in x". In this case, our current computers are grossly innacurate. 0.15 is very far away from infinity. In fact, all advantages not in the form of 0.00 or "mate in x" are grossly innacurate."

Tell me somthing I don't know. It's been repeted that there is no "1.00+" and it's only "mate in x" or 0.00 many times, I don't see how this blows my argument to infinity.

"3. Grandmasters, Masters, and Experts, all of whom know infinitely more about chess than you or I, are almost universally of the opinion that chess is theoretically a draw."

So what? Logic would tell you white wins.

Your way off. There are many forced wins. And I'm still talking about perfect players, if you are perfect, there is usally only one move.

I've said it before. A 1/3 pawn is worth about a tempo. With a perfect white, there is no way to take away that advantage.

"No. How many times have you seen even the top players and even top computers in the world fail to convert a supposedly large advantage into a win, even with lots of time to calculate? I don't know about you, but for me, the answer is "lots".Technically, having a king and knight against a king should count as a +3 advantage, 20 times your stated opening advantage, and still insufficient for a win. Certain fortress positions give similar evaluations (+3 or more) against the defending pieces, but it is completely impossible to beat them, and they are a draw. Why do you assume the much smaller +.15 advantage is a win, where these advantages are a draw? Computers don't know everything, especially when they've been programmed by humans."

Again, I'm talking about the 1/3 pawn. And a K+N vs. K is 0.00 or drawn. 

"Actually the statistics favour white about the same amount at every level, the only difference being that more draws happen as skill goes up. "

As said before, from 1200 to 2200 white's wins rise.

"No, they lower. People about my level probably get 50% wins, 10% draws, 40% losses as white, give or take. At master level, as you mentioned, it's about 40-30-30, so white's winning chances actually go down significantly. "

No, white's chances rise over black from 1600 to master level. At first, black has  4/5 as much chance to win as white. Then, at master level. Black only has a 3/4 chance of winning.

"Nope. The wins at any level are actually caused by tiny (or big) errors. Find me a single game where anyone wins where no mistakes are made by the losing side. You can't."

That's because the perfect game has never been played. And that's only your opinion, but you say it like a fact.

I have no idea what computer you're using, 1000 elo? lol nevermind. I'm not puting my faith in computers, I'm putting it the 1/3 pawn.

White does not have some advantage, with perfect play. That is what we are arguing here. It's quite the fallacy to use the conclusion you want to reach to back up your argument for that conclusion. Now please, reread my first post until your brain has evolved sufficiently to understand it. 

That's not what I meant. How is an advantage of 1 tempo enough to cause a win? I doubt it can.

Houdini. Beater of Rybka. Best computer engine in existence.

There is no 1/3 pawn. There is just the opinion of some masters (who are not perfect players) who say that 3 tempi is worth about a pawn. Neither an extra tempo, nor an extra pawn, nor an extra knight, nor anything except a proven win is enough to win a game.

How do you know this wonderful advantage is more than the kind of advantage which leads to a king and two knights vs king endgame? Or kind and rook pawn vs king endgame? OR king and pawn vs king where black has the opposition? Or drawing fortress position?

Advantage does not equal win. Advantage can just means that the disadvantaged player needs to play more carefully to avoid a loss.

Only because people writing the program wrote it into the code - the computer doesn't know it's a draw, it's TOLD it's a draw by the programmers. If they didn't, it would evaluate as around +3. And your 1/3 pawn does not exist.

I'm not talking about relative chances of winning, I'm talking percentage of games which end in a win for white. Clearly. Don't try to be smart by changing the goalposts.

And white's chances don't rise. White scores .55 on average in both examples. Why you think that should rise to 1 is not explained at all (because you have no reason).

It's a fact until you disprove it. It is true for every single game of chess played in history, human or machine. Find me a game where no mistakes are made, and I'll show you a draw. Or more likely, a mistake.

As you say, the perfect game has never been played. Yet somehow, because some masters consider a tempo in the opening to be worth 1/3 of a pawn and white has an average of 1/2 a tempo in advance of black, you believe yourself fit to pronounce the perfect game a win for white, when you're lower rated even than me, let alone all the grandmasters who call the game a draw, the computers which beat this humans and still predominantly draw against each other.

You're 12. You can't afford to be this arrogant for the rest of your life. Learn when you're wrong. Like you said "it's only your opinion, but you say it like a fact".

...After spending the past five pages worth of posts arguing that computers are infallible.

When the rules of chess are changed so that you get zero for a draw or loss and one for a win, the idea of counting wins for white alone might make some sense. Until that day, it is the difference between the percentage of wins for white and the percentage of wins for black that is the relevant statistic to determine white's practical advantage. One would expect the advantage of the first move to have a greater influence on the result for stronger players, and it does.

Matthew11 always claims that draws happen less frequently than White wins in master play.  Let's take a look a some recent World Championship matches to see if he's right. 

2010: 4 White wins, 1 Black win, 7 draws

2008: 2 White wins, 2 Black wins, 7 draws

2006: 6 White wins, 3 Black wins, 7 draws

2004: 2 White wins, 2 Black wins, 10 draws

Seems there are way more draws than White wins.  While there may be more white wins in master play, that is simply not the case in grandmaster play.  This is because GMs are better and so they do not make as many mistakes as ordinary masters.  So if better correlates to an increase in draws, perfection correlates to a drawn game. 

"It's a fact until you disprove it...."

Hmm, what if I told you "watermelons are blue on the inside before you cut the skin.  Prove me wrong." Is that a fact?

I didn't say that. I said white wins more than black the better the players are.

Actually, at the world champion matches the white wins ratio over black's is very high with:    37% 1-0      56% 1/2-1/2     7% 0-1

As I was saying, the better the players are, the more white wins over black until at world champion level, black only wins 1 in 14 of games.

Enough said.

A tiny advantage doesn't mean white "wins by force." If you're in a K+P vs. K situation with the king blockading the pawn, the side with the pawn could be said to have a "slight advantage," but it's still a draw.

Honestly, though, who cares? even ifa computer solves chess, there's no way ANY human being could keep the millions of possibilities in his brain. Human chess will always contain flaws, which is certainly a good thing.

No, it is the percentage of wins plus half the percentage of draws for each side which determines the advantage. Otherwise white winning 10% and drawing 90% would give white an infinity advantage.