Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: ==> game tree of perfect chess

Author: Tony Werten

Date: 02:50:23 01/09/03

Go up one level in this thread


On January 08, 2003 at 10:58:11, Vincent Diepeveen wrote:

>On January 08, 2003 at 09:33:04, S. Loinjak wrote:
>
>>I'm not sure how the complete game theoretical tree of chess looks like but I've
>>got the folowing ideas about it.
>
>>From correspondence chess I know that high search depths (e.g 24 ply in middle
>>game with still 20 men on the board) enable you to start optically very very
>>risky looking attacks where the initiative gain can compensate a considerable
>>amount of material over a long time.
>
>that is however not near the theoretical number of board positions. in my own
>games (i am FIDE master means >= 2300 FIDE : like 2500 USCF) i also can
>sometimes make up for big material. However in perfect chess that is far more
>complicated with exception of reaching strategical goals from opening. We do not
>talk however about more than the material that's on the board. After a few
>captures of course the number of positions go down bigtime.
>
>the problem of the number of theoretic positions is that most positions are
>possible with the entire set of pieces on the board, whereas the reality is that
>you directly swap a pawn or so in most openings.
>
>If you do not swap the pawns run against each other which limits the number of
>possibilities also considerable.
>
>10^43 which is last calculated number is not near the real truth simply.
>
>there is so many nonsense positions reachable with a full board of pieces
>whereas usually they stay on the same half of the board. My pieces on my side of
>the board and the opponent his own. at the 4th and 5th rank usually both have
>some material mixed and there is very little pieces of mine inside the opponent
>positions.
>
>That limits the number of possibilities by magnitudes already. Getting down from
>10^43 to 10^38 is really a peanut with such rules which are never hard rules.
>
>the difference between optimal chess and normal chess is very similar to each
>other. The difference between normal chess and theoretic possible positions is
>*completely* different.

You can take this to an extreme. Suppose, in chess the one who gets a pawn on e4
wins. State space and complexity values are the same, yet the game would be
trivial.

Tony

>
>>Therefor I 'feel' that optimal chess might be by far different from 'normal'
>>human chess. Maybe it'll be full of extreme attacks like Nezhmetdinov used to
>>play (the one who outcombined M. Tal in his [Tals!] best days - even Tal was
>>proud of those losses). Of course the main variation could look very
>>conventional and lifeless as maybe both colors are forced to act extremely
>>prophylactically to avoid a 'perfect' attact.
>>
>>Therefor I could imagine (but I'm not sure about it) that there are lines in the
>>perfect chess tree (containing the main variation(s) and at least one refutation
>>[not necessarily the strongest one] for each suboptimal move) which are highly
>>material imbalanced over a long time until mate or draw is forced.
>>
>>
>>Sini



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.