Game Theory - Perfect Strategy?

Sort:
Avatar of Elubas

Double standards: They exist in chess, too.

Avatar of Elubas
0110001101101000 wrote:
Bur_Oak wrote:
Memorizing [the solution to chess] would likely be vastly more difficult than memorizing MCO.

I know this is 5 year old, but holy crap. This is like saying the sun is likely bigger than a tennis ball.

I guess his thought might have been that there is just one big solution line and you can just go down that and ignore everything else. Though that gets broken down quickly since the solution must cover every reply :)

Avatar of u0110001101101000

Whenever it's the nameless faceless "them" vs the other "them" then yeah, there tends to be crazy contradictions wink.png

Avatar of u0110001101101000
Elubas wrote:
0110001101101000 wrote:
Bur_Oak wrote:
Memorizing [the solution to chess] would likely be vastly more difficult than memorizing MCO.

I know this is 5 year old, but holy crap. This is like saying the sun is likely bigger than a tennis ball.

I guess his thought might have been that there is just one big solution line and you can just go down that and ignore everything else. Though that gets broken down quickly since the solution must cover every reply :)

Yeah, but anyone who's ever tried to learn... just about anything in chess... should realize that knowing 1 line amounts to nothing as soon as the opponent deviates.

I realize not everyone has though.

Avatar of bbeltkyle89
0110001101101000 wrote:
bbeltkyle89 wrote:
philidor_position wrote:

It's obviously a draw with perfect play, but how does that make it exactly the same with tic tac toe? Why the rush to reach so quick conclusions? Basketball is exactly the same as football, except that it's completely different.

Is it so obvious? i like to hope that black has the ultimate zugzwang, where any move white makes is a "mistake".

I've said this a few times in the past, but I like the ideas, so I'll repeat them.

What can we say observationally about zugzwang? The chance of zugzwang is inversely proportional to the number of pieces on the board and the number of legal moves. We see in the starting position maximum pieces, and many legal moves.

What can we say observationally about drawn positions? They're proportional to symmetry and inversely proportional to threats/forcing moves (or we could say contact between the two forces). Again, we see in the starting position perfect symmetry and no threats (no contact).

So based on experience, and at a glance, it certainly seem chess is a draw.

Honestly, yes i think it might "seem" that chess is a draw.  I just dont think it is obvious....or worth finding out for that matter.

Avatar of Elubas

If you watch top level chess and how hard it is for either side to take any control of the game, you'll much better understand why the opinion of chess being a draw is so strong and popular. And as I said before, the analogy to resigning applies. Let's say you lose a queen in the opening for no compensation. It's "clearly lost," but it would be impossible to prove just by listing all possible variations, because there are too many. But that's not a reason to have serious doubt about the theoretical result.

Avatar of u0110001101101000

Yeah, or if you realize what the drawing margin for practically 100% of endgames is... you can't just be a little ahead and win a game of chess (with best play).

Avatar of Elroch

There is sort of a race with imperfect knowledge (always the case until chess is solved). One side has a slight practical advantage (eg as assessed by a top engine, or traditionally by GM assessment), which gives it a better chance to get the opponent into a corner where it eventually turns out to be losing. But there is a contrasting effect that small advantages become increasingly likely to turn into dead draws in endgame positions because effectively there is a time limit kicking in - the practical number of moves in which an advantage might be turned into a win.

Encouragement for the majority view that chess is a draw is given by the fact that the large majority of opening lines reach a position which is considered equal, rather than leaving white with an advantage. This suggests that there is a tendency for small practical advantages present early on in most high quality openings to dissipate as black has sufficient resources to hold the position.

Avatar of u0110001101101000
Elroch wrote:

Small advantages become increasingly likely to turn into dead draws in endgame positions because . . . the practical number of moves in which an advantage might be turned into a win.

What? I read this a few times and I don't understand what you're saying. Are you talking about the 50 move rule?

When I say a small advantage isn't enough to win a game, I mean like being a pawn down but the tablebase says draw. (And a pawn down is of course massive compared to usual advantages.)

Avatar of Elroch

No, I am referring to the situation where a win is beyond your horizon but you know you have a practical advantage, and it turns out that as both players play apparently accurately, the advantage grows eventually into a winning one. This phenomenon happens with both strong players and computers.

The practical chance for an advantage to have time to "grow" into a win is more limited the more towards the endgame you are.

Avatar of xman720
JG27Pyth wrote:

Chess may be a devilishly difficult endgame study composed by God... "White is in Zugzwang! 0-1 "

Avatar of TheIronDuke
0110001101101000 wrote:
bbeltkyle89 wrote:
philidor_position wrote:

It's obviously a draw with perfect play, but how does that make it exactly the same with tic tac toe? Why the rush to reach so quick conclusions? Basketball is exactly the same as football, except that it's completely different.

Is it so obvious? i like to hope that black has the ultimate zugzwang, where any move white makes is a "mistake".

I've said this a few times in the past, but I like the ideas, so I'll repeat them.

What can we say observationally about zugzwang? The chance of zugzwang is inversely proportional to the number of pieces on the board and the number of legal moves. We see in the starting position maximum pieces, and many legal moves.

What can we say observationally about drawn positions? They're proportional to symmetry and inversely proportional to threats/forcing moves (or we could say contact between the two forces). Again, we see in the starting position perfect symmetry and no threats (no contact).

So based on experience, and at a glance, it certainly seem chess is a draw.

 

Chess doesn't have perfect symmetry, the position of the king and queen are switched for black and white, which is why if you both castle kingside you are both on the same side of the board, instead of opposite.

Avatar of bbeltkyle89
TheIronDuke wrote:
0110001101101000 wrote:
bbeltkyle89 wrote:
philidor_position wrote:

It's obviously a draw with perfect play, but how does that make it exactly the same with tic tac toe? Why the rush to reach so quick conclusions? Basketball is exactly the same as football, except that it's completely different.

Is it so obvious? i like to hope that black has the ultimate zugzwang, where any move white makes is a "mistake".

I've said this a few times in the past, but I like the ideas, so I'll repeat them.

What can we say observationally about zugzwang? The chance of zugzwang is inversely proportional to the number of pieces on the board and the number of legal moves. We see in the starting position maximum pieces, and many legal moves.

What can we say observationally about drawn positions? They're proportional to symmetry and inversely proportional to threats/forcing moves (or we could say contact between the two forces). Again, we see in the starting position perfect symmetry and no threats (no contact).

So based on experience, and at a glance, it certainly seem chess is a draw.

 

Chess doesn't have perfect symmetry, the position of the king and queen are switched for black and white, which is why if you both castle kingside you are both on the same side of the board, instead of opposite.

wat? 

https://en.wikipedia.org/wiki/Reflection_symmetry

Avatar of u0110001101101000
TheIronDuke wrote:

 Chess doesn't have perfect symmetry, the position of the king and queen are switched for black and white, which is why if you both castle kingside you are both on the same side of the board, instead of opposite.

As far as the pieces, the board has horizontal symmetry in the initial position.

In any case, symmetry itself means nothing... as a premise I'm using the observation that positions like these are equal:


So  even if you find a definition of symmetry for which the statement is not true, it's only an argument against my terminology.

Avatar of u0110001101101000
Elroch wrote:

No, I am referring to the situation where a win is beyond your horizon but you know you have a practical advantage, and it turns out that as both players play apparently accurately, the advantage grows eventually into a winning one. This phenomenon happens with both strong players and computers.

The practical chance for an advantage to have time to "grow" into a win is more limited the more towards the endgame you are.

Interesting idea, I see what you mean now.

Avatar of Elroch

There is generally a trade-off between computing time and storage capacity in computation. For example, the tablebase of chess is a hypothetical database (impractical on account of its astronomical size) which allows instant determination of the evaluation of all possible moves in a position.

At the other extreme it is possible (easy, in fact)  to write a tiny program that uses a small amount of memory which could evaluate all the possible moves in any position, with the only problem being that the calculation for a position usually takes longer than the age of the Universe.

Avatar of efintushel

The history of the (moves of) a game is irrelevant in adjudicating any given position.

Chess is clearly a two-person zero sum game, so it has an optimum strategy--I think Von Neumann proved this with an analogy to spherical geometry.

Nobody knows if it is optimally a win for white or for black, or a draw, and I don't think players' consensus matters on this point. For example, when the tables for a crertain set of endgames was actually computed some years ago, there were found endings previously universally considered draws that, in fact, were wins--though the winning sequence was tactically incomprehensible to any human.

Avatar of efintushel

By the way, someone upstream gave the wrong number of opening moves in tic-tac-toe. Because of the symmetries, there are only three: corner, middle, or side.  (I think the earlier poster mistakenly said that there were six!)

Avatar of Elroch
efintushel wrote:

The history of the (moves of) a game is irrelevant in adjudicating any given position.

While the rest of your post was correct, this is an error. The value of a position in competition rules chess (eg FIDE, USCF) can depend crucially on:

  1. the ply count since the last irreversible move
    and
  2. which positions have been previously visited twice

To understand the latter, consider the case where there is a single path to a win, but that path goes through a position that has been visited twice.

 

Avatar of tygxc

@107

"The history of the (moves of) a game is irrelevant in adjudicating any given position.
this is definitely an error." ++ No, it is not an error.

"The value of a position in competition rules chess (eg FIDE, USCF) can depend crucially on:
the ply count since the last irreversible move and which positions have been previously visited twice" ++ Visiting the same position twice can always be avoided. The 50-moves rule is never triggered with perfect strategy i.e. optimal play from the initial position.