Author: Terry McCracken
Date: 12:44:46 02/25/01
Go up one level in this thread
On February 25, 2001 at 15:10:50, Jorge Pichard wrote: >On February 25, 2001 at 12:29:57, Ed Schröder wrote: > >>On February 25, 2001 at 12:18:18, Uri Blass wrote: >> >>>On February 25, 2001 at 11:56:08, Ed Schröder wrote: >>> >>>>On February 25, 2001 at 11:04:52, Amicitia Stone wrote: >>>> >>>>>This is just the way I see things: >>>>> >>>>>Most impressive result: Yace 0.99x 5th place (Wonderful) >>>>>runner-up: Comet B.31 7th place (Still had 5.0/9 just like Yace) >>>> >>>>Very good indeed.... >>>> >>>> >>>>>Most disappointing result: Nimzo 8.0 11th place (What happend?!) >>>>>runner-up: P. ConNerS 10th place (How many processors does this thing need to >>>>>win?!) I expected more from it.... >>>>> >>>>>The team I feel most happy for: Deep Shredder 5.0 (Congrats!) >>>>>runner-up: Holmes 0.74 (Nice try! I hope it was a good experience!) >>>> >>>>The pattern and story continues.... >>>> >>>> >>>>>The team I feel bad for: Deep Fritz Paderborn (Soooo close! 2nd ain't bad!) >>>>>runner-up: Nimzo 8.0 (obvious reasons. It wasn't even going to enter the >>>>>tounament. I bet they regret it.) >>>> >>>> >>>>Same pattern... >>>> >>>>Put it into the autoplayer and it wins every long match. >>> Starting >>>>slowly but in the end the learner gets the other computer opponent. >>>> >>> >>>I do not see it in the ssdf match against Deep Fritz. >> >>Then have a look again. >> >>Ed >> >> >>>>Manual tournaments and autoplayer tournaments are different stories. >>>>Wish it was different and it would give an equal pattern. >>>> >>>>Ed >>> >>>I think that a bigger book may help for autoplayer tournament because the >>>program with the bigger book can find after enough games an opening when the >>>opponent has no good answer to it. >>> >>>I think that it may be a good idea to use some fixed positions similiar to the >>>nunn match for autoplayer games(I do not suggest the nunn match position because >>>they are known positions and it is better to use positions after 10-15 moves >>>from grandmasters games). >>> >>>Uri > >Simply test programs without book opening ! > >Pichard. I know Alan Tomalty, a Canadian chessplayer, and computer programm tester, has suggested that chess programms should be tested with books and and "thinking" turned off. He mentions this in The Chess Federation of Canada's magazine En Passant back in thelate 1980's. I'd really have to dig around to find the issue. But I remeber him saying this and it does have a point, when you want to find the _True_ strength of a given chess programm/engine. However today there are other factors to concider as was pointed out about _Book Learning_ etc. However, it would still work to test the engine alone. Terry McCracken
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.