Opening: Can White Get A Sustainable Advantage?

[towforce]

At a chess tournament, every entrant has to ensure that their opening moves will not be known to their opponents. In the 1990s, there were cases of opening books in commercial engines being written against other specific commercial engines as far as move 19 to ensure that they'd win in ratings games.

Today, I doubt that this happens when opening books are very large and highly engineered to avoid moves that lead to positions that normally lose.

My question relates to whether chess is a win for white or a draw, and the question is: as opening books get better, is there any sign of an emerging advantage for white when the books are left?

Disclosure: I think that chess is a draw, so I expect the answer to be that as the arms race continues, it is not resulting in any noticeable improvement in white's evaluation when the game reaches the end of the book.

[ernest]

My question relates to whether chess is a win for white or a draw,

As of today, loss for white cannot theoretically be excluded (for instance, because of zugzwangs)...

[towforce]

My question relates to whether chess is a win for white or a draw,

As of today, loss for white cannot theoretically be excluded (for instance, because of zugzwangs)...

While it's not possible to PROVE that chess is not a win for black, IMO it can be safely discounted. With such a large game tree, the probability that white would be unable to forcibly avoid an unfavourable zugzwang becomes vanishingly small (like the probability of tossing n heads and no tails with a fair coin when n>100).

Again IMO, the only way that chess can possibly be a win is if white is able to forcibly sustain an advantage that is sufficiently large to be maintainable throughout the rest of the game. If this is the case, then the first place it would be expected to appear would be in the form of white leaving the opening with an advantage in games between computers with large, well-engineered opening books.

[duncan]

While it's not possible to PROVE that chess is not a win for black, IMO it can be safely discounted. .

if black gives up a knight at the beginning, it is a win for white. do you know if it can be proven though, or merely safely discounted that white would not win and would be either a draw or even more ludicrously a win for black because of zugzwangs.

duncan

[Uri Blass]

My question relates to whether chess is a win for white or a draw,

As of today, loss for white cannot theoretically be excluded (for instance, because of zugzwangs)...

While it's not possible to PROVE that chess is not a win for black, IMO it can be safely discounted. With such a large game tree, the probability that white would be unable to forcibly avoid an unfavourable zugzwang becomes vanishingly small (like the probability of tossing n heads and no tails with a fair coin when n>100).

Again IMO, the only way that chess can possibly be a win is if white is able to forcibly sustain an advantage that is sufficiently large to be maintainable throughout the rest of the game. If this is the case, then the first place it would be expected to appear would be in the form of white leaving the opening with an advantage in games between computers with large, well-engineered opening books.

The experience in correspondence games when using engines is allowed suggest that chess is a draw.

Most games between strong correspondence players who usechess engines are draws and not win for white.

[towforce]

The experience in correspondence games when using engines is allowed suggest that chess is a draw. Most games between strong correspondence players who use chess engines are draws and not win for white.

Thanks - that was very helpful. I have added that snippet of information to my archive.

[towforce]

if black gives up a knight at the beginning, it is a win for white. do you know if it can be proven though, or merely safely discounted that white would not win and would be either a draw or even more ludicrously a win for black because of zugzwangs.

It seems to me that if that could be PROVEN, we'd be over halfway to proving the outcome of chess.