Chess, a solved game in the future?

Actually FatBot’s reply was “arrrgh. You win Again
Sorry, but I felt that was integral to the thread

Zugswang is where one player to move must weaken their position and it affects the result (usually turning a ‘draw’ into a ‘loss’).
If the attacking player has waiting moves, then it doesn’t matter if it is their move - they make a waiting move and zugswang results.
If the attacking player has no waiting moves, then it is mutual zugswang: the player to move must weaken their position.

A ‘squeeze’ is normally used to mean when one player has a space advantage and can build up pressure, whilst stopping counterplay.

(I’m quoting GM John Nunn - who is your writer?!)

This is theory, not practice.

And I agree that it’s hard to see that it would make a difference. But I was just pointing out that it COULD make a difference. If analysis shows that black has a forced win every time, white passing the first move does not automatically have a forced win every time, because the board isn’t symmetrical – white begins with his queen to the left of his king from his point of view, and black begins with his queen to the right of his king from black’s point of view.

Yes it is a theory. So is ‘Intelligent Design’. :frowning: However there is no evidence for either.

Based on my extensive experience of international chess, the only ways that Black wins at chess are:

  • White makes the last mistake
  • White plays a risky variation (because a draw is no good to him)
  • White goes in for extreme complications, which eventually turn out badly
  • Black is a much better player than White

All serious chess tournaments go to great lengths to give players equal Whites and Blacks. As I said, there is a tiebreak system where Black effectively gets draw odds.
Here is a compilation of 2 million games, showing that White has a clear advantage.

There is no opening for Balck that achieves even 50% results!

What? No, it’s not possible for that to make a difference. It’s just a horizontal flip of the initial configuration, and beyond that aspect of the initial set-up, all of the rules of the game are symmetrical with respect to left vs. right; black can just watch himself playing in the mirror, if he likes, and pretend the configuration he started with was the same as the one white usually starts with.

I realize that white has the advantage according to all known theory about the game, really I do.

What I was responding to was the notion that IF (if) it somehow turned out that black had a forced win, white passing its first move would “turn the tables” so to speak. I thought maybe that was not the case, because of the small difference in position.

Read again the statement I originally responded to:

My response was wondering whether the transposed king and queen might have an affect on this perfect line, and whether it might make a difference if black tried to play it instead of white. It had nothing to do with the likelihood of this being true, only the possibility. Likewise, it has nothing to do with the current data overwhelmingly giving the advantage to white. If you’d like to comment on the technical aspects of why black’s queen on the right can’t have an affect (as I see on preview that Indistinguishable has done), I will surely listen to you, but this has nothing to do with the statistics compiled so far.

This is fairly convincing. It was odd to wrap my head around though. Thanks.

Everyone agrees that the only way for the initial configuration to count as zugzwang is if the full game tree of chess is such that the second player wins in “perfect play” (i.e., there is a strategy for the second player which forces a win). The discussion of that somewhat humorous notion launched from that hypothetical possibility.

I think we’ll all also defer to your knowledge that the general sentiment among chess experts is that it is extraordinarily unlikely that chess is game-theoretically a win for black, despite there being no mathematical proof or even very forceful mathematical argument for this (the same way the general sentiment among computer scientists is that it is extraordinarily unlikely that P = NP, despite there being no mathematical proof or even very forceful mathematical argument for this). But all the same… well, I’d finish my post, but I see on preview that Garfield has begun to say what I was going to: we’re talking about the “IF”, and why not? Given that it has not been conclusively demonstrated that black lacks the initial game-theoretic advantage in chess, there is non-null value (even if only amusement value) in mentioning the open possibility of initial zugzwang.

Remember that debate we were caught up in about ultra-accurate predictors of human behavior?

There’s a reason I so passionately defended their possible existence… :slight_smile:

Count me in, too. Whether it is possible seems to be undeterminable, at least for now and among us. That’s why I called it an epistemic possibility, which I think is fair.

Oh, yeah, when I saw that, I felt that was a very good way to phrase the situation.

OK, I accept that I can’t disprove your theoretical premise with my mass of statistics.
Would you accept either of these arguments?

A)

  1. The only way the game can be a win for Black with perfect play is if Black can eventually force zugswang.
  2. Zugswang is defined as being in a position where any move weakens your position.
  3. At the start of the game, many move choices improve White’s position (I don’t like 1. a4, for example!)
  4. Therefore White will not be forced into zugswang and the game is not a win for Black.

B)

  1. The only way the game can be a win for Black with perfect play is if Black can eventually force zugswang.
  2. Zugswang is defined as being in a position where any move weakens your position.
  3. So (if chess is a win for Black) when White makes his first move, he weakens his position.
  4. Therefore Black should pass (I know he can’t, but this is a theoretical discusssion) and let White move agian, since White will further weaken his position.
    Sadly this strategy leads to a rapid loss for Black…
    P.S.
    For interest, there are 3 non-symmetries in the starting position:

a) White moves first.
This is a real advantage, giving about 5% in practice.

b) The King and Queen positions are not symmetrical.
No reason to assume it matters (as Indistinguishable posted).

c) There is a White square on the right.
This shouldn’t matter either - but there was a small sample of Rook + Bishop v Rook endings (typically drawn, but difficult to defend) which showed that players had slightly better success winning with a bishop on White squares. The reason might well be that endings books showing how to play this ending invariably put the bishop on a White square for better visibility…

Bruce Pandolfini, Pandolfini’s Endgame Course:

Note that Pandolfini himself does not use the words that way, but he asserts that other writers do, and that’s good enough for me. Wikipedia makes the same assertion.

Please don’t get the wrong impression. Your masses of statistics do give some reason to believe you are correct. On this one level, I think we all mostly agree with you, that black probably doesn’t win with perfect play. As Garfield said above, we acknowledge this, we defer to your expertise here. But those statistics about some far-less-than-exhaustive number of games that happen to have been historically played don’t provide anything like a mathematical argument against “the game tree of chess such that black can force a win”; the possibility, though we may all feel, and be justified in feeling, it unlikely, remains “open” all the same.

Not in this context, no. In the full relevant game-theoretic sense, the only possible “values” positions can have are “The game is a forcible win for white now”, “The game is a forcible win for black now”, and “Both players can force non-loss now, so perfect play will result in a stalemate now”. Keeping that in mind, step 3. is both an attempt to beg the question, and, in fact, false:

To know the actual value of white’s position after such moves in the full relevant game-theoretic sense, we would have to actually provide a mathematical argument for the result of perfect play after such moves.

However, we may as well note, in the game-theoretic sense, on white’s turn, at any moment, the value of his current position is the maximum of the values of all the positions he can move to (because, after all, he can, uh, move to them). In this sense, no one can ever make a move which improves his position; he can either maintain it or make a blunder and drop it. So the premise in step 3 is actually false, on that account.

Finally, though not terribly importantly,

Well, as you say, it doesn’t really matter. It matters as a(n amusing) fact about human nature, but in terms of the mathematical question, “Is the game tree of chess such that black can force a win?”, it is entirely irrelevant.

OK, good research! :slight_smile:

But from your Wikipedia article (bolding mine):

'There are three types of chess positions:

  • both sides would benefit by it being their turn to move
  • only one player would be at a disadvantage if it were his turn to move
  • both players would be at a disadvantage if it were their turn to move.

The great majority of positions are of the first type.

In chess literature, most writers call positions of the second type zugzwang, and the third type reciprocal zugzwang or mutual zugzwang. Some writers call the second type a squeeze and the third type zugzwang.’

Given that GM John Nunn, for example, uses ‘mutual zugswang’ and that Wikipedia quotes David Hooper and Ken Whyld (much weaker players) as using ‘squeeze’, I’m sticking with my vocabulary!

Oh, you edited your post after I replied to it (I’m always afraid I’m going to do this myself; I like to posit the mythology, like the tale about some tribe or another and their fear of cameras, that if someone quotes my post in a transient state of editing, showing to posterity the words I later chose to modify, they’ve captured my soul). I replied to argument A, but let’s touch on argument B as well:

Step 4 is fallacious. We can’t conclude that subsequent moves for white will weaken his position just from the fact that white’s first move will weaken his position.

Yes, I accept your precise statement (and apologise if I was ‘showing off’ - I found the OP’s unsupported assertions e.g. that the Najdorf was ‘basically the most perfect 5 moves’ irritating :rolleyes: ).
I have not proved my case conclusively.

As a discussion point though, isn’t this similar to the Evolution v ID argument? In both cases there is a lot of data for one theory and none for the other.

Oh, indeed. I just thought you’d find it interesting. :slight_smile:

Clearly I was not precise enough in my statement, since you’re both saying I’m wrong, but I don’t see the major difference between my statement and your (more detailed) explanations. As far as I can see my major error was failing to specify “or draw” after “win” (in the case that that’s the best outcome for a player).

If “solving” a game means that you can predict the outcome of the game, then you must know what perfect play is. If you know what perfect play is, then the player who will win can employ that strategy to do so. And, certainly, anyone playing against a player with the best strategy only has a chance to win if they, too, play perfectly. Isn’t that pretty much what I said? Or do I misunderstand. psychonaut, could you provide a trivial example for me?

Sorry about the edit - it’s midnight where I am (and I’m playing an Internet game too :eek: )

I’m interested in your statement ‘We can’t conclude that subsequent moves for white will weaken his position just from the fact that white’s first move will weaken his position’.

OK, White makes an initial move that weakens his position. However he now threatens to improve his position.
Black (who we are assuming has a forced win) swiftly makes a move that prevents the improvement.

Now I know of games where this applies. But I still feel strongly that such a case can’t apply in chess from the starting position. (I know, this isn’t proof!).

You’ve got me thinking - if Black did have a forced win against any move by White, wouldn’t there be at least one example in the last 400 years of chess of Black winning by just good play?
Because there are plenty of examples of White gradually outplaying Black, but none of the other way…

I suppose this is like asking for one miracle to prove that God might exist. :confused:

This is already not true. I gave the example of Hex; we know that the first player in Hex can force a win (on any board size) but we don’t actually know how to do so, beyond very small boards. We don’t even know what first move would be appropriate for him to take. What we have is a non-constructive existence proof.

Well… if the first player can force a win, then the second player can only win if the first player blunders and deviates from perfect play. But the second player doesn’t have to play perfectly for the first player to blunder. And if the first player doesn’t blunder, then the second player has no chance to win, period.