Author: Christophe Theron
Date: 16:48:42 02/16/00
Go up one level in this thread
On February 16, 2000 at 12:38:26, Albert Silver wrote:
>On February 16, 2000 at 04:02:16, Tom Kerrigan wrote:
>
>>I had an idea tonight: I would watch Genius (on my Palm III) crush TSCP (on my
>>Celeron/400).
>>
>>Well, that idea backfired.
>>
>>I played two games between the programs. Genius was set to 1 sec/move for both
>>games. TSCP was set to 5 sec/move for the first game and it absolutely crushed
>>Genius. I figured the hardware advantage was way too big, so I set TSCP to 4 ply
>>and played another game. TSCP crushed Genius again.
>>
>>Then a friend of mine came to visit. He's rated ~1500 blitz on some chess
>>servers. I asked him to play Genius. I wouldn't say he crushed Genius, but he
>>did win after 30 or 40 moves. And he was playing faster than Genius, too.
>>
>>I watched all 3 games, and Genius was making some very weird moves. It seemed to
>>cripple its own pawn structure on purpose, and it kept moving its pieces to the
>>corners.
>>
>>Comments???
>>
>>-Tom
>
>At 1 second a move I am not surprised. The program is supposed to be rated at
>2000 or so at 40 moves in 2h. The program was already quite selective, and
>didn't see all that far anyhow, so I don't know whether it will be seeing more
>than a ply or 2.
>
> Albert Silver
At 1s/move, it computes typically 3 to 5 plies ahead. Sometimes 6 plies.
Impressive. But the search is very selective, I assume it misses a lot of subtle
things.
And I notice an obvious weakness: it does not extend checks, so checks-based
combinations are often too deep for it.
I don't want to say I don't like CG for the Palm: this program is great and
certainly the best chess program you can run on your Palm.
Christophe
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.