Author: Marc van Hal
Date: 14:07:06 06/30/03
Go up one level in this thread
On June 30, 2003 at 02:28:53, Uri Blass wrote: >On June 30, 2003 at 01:35:37, Johan de Koning wrote: > >>On June 29, 2003 at 19:07:49, Uri Blass wrote: >> >>>On June 29, 2003 at 18:32:09, Vincent Diepeveen wrote: >>> >>>>On June 29, 2003 at 09:57:23, Jorge Pichard wrote: >>>> >>>>>On June 29, 2003 at 09:35:50, Sherry Washington wrote: >>>>> >>>>>>Would that be Shredder7.04, i am reffering to TOP programs >>>>> >>>>>I have not seen a better tactical program than CM9000 yet. >>>> >>>>that thing doesn't even run at 1 and a half processor and won't show up at world >>>>champs 2003 even. Very weak of the ubisoft company. >>> >>>1)The poster did not say that he has more than one processor to use. >>>I think that most buyers do not have more than one processor to use. >>> >>>2)There is still time to decide about going to the world champs 2003 and Johan >>>may change it's mind and go >> ^^^^ >>According to Bruce's spelling page that should be "its". :-D >> >>(And no, it hasn't changed its mind since 1995, and it's not planning on >>changing it yet.) >> >>... Johan > >Sorry,I think too much about computers. >The right word is his > >Uri A lot better would be to say the post started which program is currently strongest in tactics Ends up in which program might win the World Championship I guess you all still have something to learn about chess. Because chess is not only tactics But always about strategy. Chess programs still come some what short in this part and This is what makes world championship computer chess still nothing but a gamble With some knowledge about the weakness of the opponent And tweaked versions against this weakness. But not really based on real chess knowledge. Perhaps someone should look at the brain of a GM and how it comes he eventually finds tactics Many times this begins with very simple tactics (The end position), which are build up to a large more complicated variation. And for him well known patterns. In which these tactics a cure. Make the program at first make the program look for the same kind of tactics And only then start looking for the next variation. (It does not have to be move choice number one but it should not be dropped in search. So it might be candidate number4.) And only drop the line later when it turns out to be wrong. Tactics might be increased this way But also strategy. Now computers still put some candidate moves directly in the recycled bin And see them as not worth looking at. It would be a kind of super heuristics. Even when the program would perform bad in comp comp play There would be a place on the market for it. Imagine what it does when it also performs great in comp comp games. Ok I admit I also make an other post from this original post But I think it is more interesting to look at ideas on how to make programs even stronger in tactics and strategy. And on how to eventually cross the borderlines. Then to look at which program performs best in the world championship 2003. Marc
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.