Can chess be solved (in part)?

[ozziejoe]

I understand that we are a long way off finding the right response for every possible move in chess.

What if we set as our task to only partially solve chess. That is, instead of looking for solutions to every possible first move, we only look for solutions to 1. e4. We then have 12 legal moves for black. in response to each legal black move, we again to find only one white move. You could do this till the end of the game potentially? If you get to the end, and there is a loss for white, you would back up in the tree and try another line.

This same process could be done for the black side?



Is this possible? Could we restrict the range of possibilities sufficiently that chess could be "virtually solved" given a particular opening. You would not have to find every solution...

best
J

[Sylwy]

"Can chess be solved ? " [Joseph/02.08.2008 aJCH]

YES = 32 men TBS.

They rest three problems to be solved :

1.the time to compute them;
2.the hoarding capacity;
3.the practical capacities to utilize a such bulk of informations .


Good luck !


Regards,
Sylwy

[Tony Thomas]

I see 20 legal moves for black after e4.

[Ovyron]

What if we set as our task to only partially solve chess. That is, instead of looking for solutions to every possible first move, we only look for solutions to 1. e4. We then have 12 legal moves for black. in response to each legal black move, we again to find only one white move. You could do this till the end of the game potentially? If you get to the end, and there is a loss for white, you would back up in the tree and try another line.

This same process could be done for the black side?

I tried to do this some time ago, and was convinced that it's just too time consuming to be worth it. The fastest way is to solve it backwards, from 7men table bases to 8men and so on. Even that is taking too long, so don't hold your breath.

Another interesting question is, why would we want to solve chess? That would kill computer chess, as the only new interesting focus would be to know what is the optimal strategy against imperfect opposition, but all the other engines could be perfect so it would be pointless (as our imperfect engines are good enough to beat the best humans.)

[Sylwy]

Hi Tony !

The result of the perfect chess game ?
The afterlife ?..............
................................. ......................................


Hard questions !

I prefer the practical chess ( summarizing) !
And-of course- the present real life !
.................................... .............................................


Regards,
Sylwy

[gerold]

I see 20 legal moves for black after e4.

Ok that/s a start.

[S.Taylor]

I understand that we are a long way off finding the right response for every possible move in chess.

What if we set as our task to only partially solve chess. That is, instead of looking for solutions to every possible first move, we only look for solutions to 1. e4. We then have 12 legal moves for black. in response to each legal black move, we again to find only one white move. You could do this till the end of the game potentially? If you get to the end, and there is a loss for white, you would back up in the tree and try another line.

This same process could be done for the black side?



Is this possible? Could we restrict the range of possibilities sufficiently that chess could be "virtually solved" given a particular opening. You would not have to find every solution...

best
J

It would be interesting if there is a way to make a reasonable attempt to partly solve chess, but I don't think your example is any use. For a start, Black has more than 12 responses to e4, as has already been pointed out.

Are you trying to say we should only pick out 12 responses?

After every white move we should select another 12 responses to that? Never! Maybe response number 13 will win for black!

[bob]

I understand that we are a long way off finding the right response for every possible move in chess.

What if we set as our task to only partially solve chess. That is, instead of looking for solutions to every possible first move, we only look for solutions to 1. e4. We then have 12 legal moves for black. in response to each legal black move, we again to find only one white move. You could do this till the end of the game potentially? If you get to the end, and there is a loss for white, you would back up in the tree and try another line.

This same process could be done for the black side?



Is this possible? Could we restrict the range of possibilities sufficiently that chess could be "virtually solved" given a particular opening. You would not have to find every solution...

best
J

You are operating on an assumption that has not been proven. Who says that 1. h4 is not the best move for white and the only way to force a win? Until you can prove that is not the case, you can't exclude the move. And we have no hope of disproving that hypothesis for thousands of years yet. If you go deep enough into the game, you can already prove whether it is a win, draw or loss. But that doesn't help when you back up a few moves and are faced with choices that have not been proven yet...

[Ovyron]

Who says that 1. h4 is not the best move for white and the only way to force a win? Until you can prove that is not the case, you can't exclude the move.

Joseph's method would be useful to proving that Black can't win the game, and with the correct implementation, you only need to search about 5% of the chess positions you would need otherwise. Unfortunately, since they increase exponentially, it hardly matters at all.

[tano-urayoan]

This is nonsense. If white loses we go back and search for a better move so we have two results draw or white win, which one will be valid.If we find a draw we finish the experiment? But if white could win with another move? Which will demonstrate the solveness (do this word exist) of chess.
For black first move he has 20 possible responses, for his second move if is first move was a pawn he has 18 or 19 moves, if a knight move (Nc6 or Nf6) he has 21 more moves at his disposal etc. You get the point.