Author: Uri Blass
Date: 11:45:24 02/08/01
Go up one level in this thread
On February 08, 2001 at 14:01:07, Jean-Bernard Alemanni wrote:
>On February 08, 2001 at 06:17:09, martin fierz wrote:
>
>>On February 08, 2001 at 05:26:40, Uri Blass wrote:
>>
>>>On February 08, 2001 at 04:24:11, David Blackman wrote:
>>>
>>>>On February 07, 2001 at 16:41:28, Tanya Deborah wrote:
>>>>
>>>>>
>>>>>
>>>>>Hi!
>>>>>
>>>>>I am playing a new match in checkers between the 2 strongest Spanish checkers
>
>Thank you from one of "strongest" programs author !!!
>I would say that Windamas is stronger than Dynamo, but maybe I am a little
>partial :o)
>
>
>>>>>programs of the world...
>>>>
>>>>Just curious, is "Spanish checkers" the same game as "Polish Draughts",
>>>>"International Draughts", "Damen" etc?
>
>
>Spanish checkers is different from Polish. It is played on a 8x8 board, with
>double corner on left (different from american) and with flying king.
>
>>
>>there are about a zillion different rules for checkers - a nice overview can be
>>found on
>>
>>http://www.triplejump.net/rules.shtml
>
>There is another good site too, my site, where you can have explanations about
>main rules played in the world, map of the world to know what is played and
>where it is played, and where you can download programs in 9 of the main
>variations, and experiment them.
>Tanya loaded Windamas 3D there I think.
>the adress : http://perso.wanadoo.fr/alemanni
>
>>
>>>>According to people who have tried, it is a bit harder to
>>>>write a strong program
>>>>for it than for chess.
>>>
>
>>>I think that the opposite is truth.
>
>Both are probably true. I think that it is more difficult to write a good
>checkers program because you find less documentation about checkers programing
>than about chess programing.
>But, of you make 1 checkers program and 1 chess program with the same
>algorithms, the chess program will be weaker if you compare it to human players.
>So, writing a STRONG chess program is more difficult than writing a STRONG
>checkers program.
>
>>>I remember that I read that chinook won against the world champion in this game
>>>before Deeper blue(I read that the result was 2:1 and 67 draws).
>>
>>it depends on the variation. the main differences are the board sizes (8x8,
>>10x10 and even 12x12), and the rule for kings. in some variations, kings move
>>like kings in chess, in others, like queens. of course the queens-variation
>>allows many more moves than the kings-variation. if you play a queens-variation
>>on a 10x10 board ('international checkers', 20 pieces each) you have *much* more
>>complexity than if you play a kings-variation on an 8x8 board (as my program
>>does, 'straight checkers'). i think vincent diepeveen's checkers program plays
>>international checkers, so he might be able to tell us what the difference in
>>branching factor is compared to chess. in straight checkers you get *lots* of
>>draws. chinook never really beat the world champion (tinsley) over the board,
>>tinsley got ill during the rematch (he won the first match) after six draws and
>>forfeited his title. he died shortly after this. chinook then won a match
>>against the world number two with a close result.
>>the whole story can be read in schaeffer's book 'one jump ahead'.
>>in straight checkers, computers are better than humans, i don't know about
>>international checkers.
>>
>
>Programs are stronger when they have few moves to analyze.
>My programs are based on the same algorithm, and, in Brazilian checkers, in
>Spanish checkers, in Pool checkers, all 8x8, I think they have the same level as
>the best human players in the world. I tried these games on VOG, and they lost
>few games (pool version never lost on more than 40 games). That is not true for
>American and Italian checkers, because I don't know very well these games, and I
>could not set the program correctly. But Martin's program, or Chinook, show that
>computers are as good as best humans.
>In International 10x10, the program has played several games against
>grandmasters and lost so much games than it won. So, its level is lower than 8x8
>programs.
>In Canadian 12x12, I can win it sometimes (about 1 game on 10) and I am a weak
>player. I think that good players can win very often. So the program is weaker
>than 10x10.
>And, finaly, in Poddavki, which is a Russian rules give away (8x8), where there
>are few good moves at each turn, the same program is far better than the best
>human players (better than in other 8x8 rules).
>So my opinion is done : the size of the board is very important for the number
>of possible moves, and very important for the level against an human player.
>I think that if somebody makes a 14x14 or a 16x16 program, most weak players
>will win against the computers.
I believe that we only need the right algorithm if we want computers to be
better than humans in 14*14 or 16*16 or n*n for bigger n's.
Computers are clearly faster than humans and the only reason that computers
cannot beat humans in every game that is not solved is the fact that humans are
stupid and did not find the right algorithms.
Uri
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.