Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Can latest supercomputers almost solve chess yet?

Author: Frank Schneider

Date: 22:46:23 11/15/03

Go up one level in this thread


On November 15, 2003 at 23:42:40, Derek Paquette wrote:

>On November 15, 2003 at 22:09:51, Mike Byrne wrote:
>
>>On November 15, 2003 at 22:02:42, stuart taylor wrote:
>>
>>>I believe there are supercomputers that are much much faster than was the IBM
>>>which beat Kasparov with Deeper Blue.
>>>If so, what are we waiting for?
>>>S.Taylor
>>
>>Even much faster computers will not solve chess.  The possibilies of chess games
>>exceeds the number of atoms in the universe and by the time it could even be
>>theoretically solved, the Sun would have burned out and there will no humans
>>living on earth.   So one can safely say , chess will never be solved (white to
>>move and win) by man using computers.
>>
>>That would not prevent a computer perhaps being the best in the world.
>
>This gives an indication of how many legal chess moves there are at each ply,
>
>2ply =  1,600
>3ply =  64,000
>4ply =  2,560,000
>5ply =  102,400,000
>6ply =  4,096,000,000
>7ply =  163,840,000,000
>8ply =  6,553,600,000,000
>9ply =  262,144,000,000,000
>10ply = 10,485,760,000,000,000
>11ply = 419,430,400,000,000,000
>12ply = 16,777,216,000,000,000,000
>13ply = 671,088,640,000,000,000,000
>14ply = 26,843,545,600,000,000,000,000
>15ply = 1,073,741,824,000,000,000,000,000
>16ply = 42,949,672,960,000,000,000,000,000
>17ply = 1,717,986,918,400,000,000,000,000,000
>18ply = 68,719,476,736,000,000,000,000,000,000
>19ply = 2,748,779,069,440,000,000,000,000,000,000
>20ply = 109,951,162,777,600,000,000,000,000,000,000
>
>
>
>DEEP JUNIOR 8 YY
>20ply = 715,580,000
>99.999999999999999999999349183781 % filtered moves
>
>SHREDDER 7.04 (strongest settings)
>15ply = 38,826,000
>
>99.999999999999996384046971797943 %   filtered moves
>
>As you can see,  your home pc computer isn't even looking close to 1% of the
>moves,

Note that you don't have to look at all moves

- alpha beta about doubles the search depth without loosing correctness

- You don't need to search the whole game tree to prove that a position
  (e.g. the starting position) is won.

But I agree that it will take a very long time until chess is solved.

Frank


>
>Deep Blue 1997 calculated 200,000,000 positions a second,
>so if it wanted to look at a TRUE 7PLY, it would take 13 minutes and 30 seconds.
>
>I agree with Mike 100%, we'll never solve chess,



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.