Author: pavel
Date: 03:57:47 11/14/00
Go up one level in this thread
On November 14, 2000 at 06:33:09, Enrique Irazoqui wrote: >On November 14, 2000 at 06:24:51, pavel wrote: > >>On November 14, 2000 at 05:49:19, Thorsten Czub wrote: >> >>>On November 13, 2000 at 21:09:49, stuart taylor wrote: >>>>I think I understand now. >>> >>>thats always very good :-))) >>> >>>>Well! that almost makes me want to cry! (together with an emotional Gambit >>>>Tiger). >>> >>>?! >>> >>>>GT does not look anything special at all according to these results. >>> >>>50% against Junior6a. >>>So how will Junior want to beat GT in Elo-points ? >>> >>> >>>>You might say, "but it's not enough games to indicate very much"! >>>> But I can tell you, that I anyway am always upset to see a program which I >>>>want to get for analysis, and main use, lose even one game, esp. to a previous >>>>generation program, since that indicates some inadequacy somewhere, compared to >>>>something else which is available. >>> >>> >>>you are upset? why ? >>> >>>>I would prefer if they were many draws, even >>>>if only the occasional win. That would look more reliably better. But now, it >>>>looks like in many ways, the old Junior6a is stronger than GT. >>> >>>junior6a gets - if it is lucky, 50% against GT, Rebel-Tiger gets better >>>results against Junior. But - IMO GT is the better program. >>> >>>i don't see why you are upset. its very normal that programs play different >>>against different opponents. >>>you have always ONE program that is a kind of fear-enemy. >>>thats normal. >>> >>>>S.Taylor >> >> >>LOL, >>I get your point. >>when GT plays against crafty, crafty becomes "punching bag" of the old paradigm. >> >>when GT plays against junior6, the results is almost 50%, your 'defense >>mechanism' comes out saying "CT is better against junior6 then GT" >> >>DUH!!! >> >>now dont tell CT is also the result of "new paradigm" >> >>isnt this supposed to be a test between the best of "old paradigm" and the best >>of "new paradigm", that i suggested? >> >>you know, you are so predictable and I am almost damm sure what you are gonna >>say next, and all I can do is kick back and laugh about it :) >> >>pavs. ;) > >Paradigms aside, the fact is that Gambit scores very well. A worst case of 50% >against Junior 6a, Hiarcs 7.32 and Deep Fritz, while beating, sometimes >smashing, all the other programs I tried is very, very good. If you add up the >exciting and speculative playing style of Gambit, I would say that it is quite >an achievement. > >Enrique yes if you put the paradigm-concept aside, then I can understand that it has an excellent playing style, though i havent bought it yet, and not likely to buy it anytime soon (spent too much money on chess programs this year :( ). I have played many games on ICC (as guest), on FICS and also on chess.net, against it. And I have to admit that games against it were more fun that playng against CT. My guess is that it's around 20-30 elo better than fritz6 or junior6. It plays on the same level or better as some of my betas but when it wins it has a reputation of crashing it's opponents. ie, on ICC xtreme (as guest) won 3 games in a row, in long time control games on a slower hardware, then GT came back with a smashing 7 games in a row win. not to mention some of it's wins were so scary, and unpredictable. for instance, xtreme lost two games, in less than 25 moves, but all along it saw itself better in position. implementing king-side attacking style on chess program is very interesting. another good example is cometB27, it plays excellent kings-gambit, thought not in the par of GT but very entertaining. pavs. ps, by the way when you have time testing gandalf let me know what you think. I am interested on your opinion about it.
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.