Author: Vincent Diepeveen
Date: 12:44:17 07/19/00
Go up one level in this thread
On July 19, 2000 at 10:32:45, Ed Schröder wrote: I completely don't believe in preprocessing. DIEP is not preprocessing at all actually. With some smart tables you can still fool many humans however and still have a program playing a lot of average moves, but preprocessing or lack of knowledge clearly isn't the way to go. To answer the question of Amir: "do you believe junior is a preprocessor?" my answer is: yes i do. Note that doing lots of alfabeta dependant stuff also looks very preprocessing, but i get the impression that junior doesn't have a qsearch at all, but just a very quick. A program that's NOT preprocessing just can't ignore the fact that you need to make moves and THEN evaluate, where the huge advantage of preprocessing is not only speed but also the possibility to do qsearch much simpler. I can't prune at alpha for example in qsearch, that's very dangerous, because capturing a pawn might evaluate into something completely different. However with a preprocessor you can very clearly say when looking at a possible to capture pawn: expected score = curscore + pawn - pawnvalue[squareofpawn] Now you can heavily prune. When having a huge eval that's actually impossible to alpha prune. Now in Juniors case i get the impression it doesn't do *anything* what looks like my qsearch. It seems to just 'estimate' it and goes on. A lot of principal variations of junior are perhaps simply completely idiotic therefore. That they're completely insane is clear. I'm not talking about the QUALITY of the lines here, but the end of the line, namely that it tries some capture moves which only a preprocessor would try, thereby missing other things. Now this can be because of a hashtable side effect, but it sure is something which i hardly see in other programs, except for some known preprocessors. >On July 19, 2000 at 08:22:18, Amir Ban wrote: > >>On July 19, 2000 at 01:11:55, Ed Schröder wrote: >> >>>On July 18, 2000 at 19:10:46, Amir Ban wrote: >>> >>>>On July 18, 2000 at 14:05:46, Jeroen Noomen wrote: >>>> >>>>>On July 18, 2000 at 09:29:12, Amir Ban wrote: >>>>> >>>>>Amir, >>>>> >>>>>I agree that Junior earned its points honestly. I also agree with most you write >>>>>about these games. Still, you don't point out anything about the losses against >>>>>Kramnik and Piket. And that was exactly what I had in mind writing this thread. >>>>>Those two games showed exactly where chess computer programs still can be >>>>>improved. And HAVE to be improved, otherwise human GM's will have good chances >>>>>to get more points next year. And they will, because they have learnt. >>>>> >>>>>IMO if you solve most of the problems about king's attacks and closed positions, >>>>>then it will be almost impossible for the strongest GM's ta beat a computer. >>>>>Because in that case they have no advantage in any type of position anymore. But >>>>>in 2000 there is still not much to be done when a clever player manages to block >>>>>the position or start a slow attack: The programs do not know about this and >>>>>only human mistakes will save them. >>>>> >>>>>So the crucial question is: When will one of the leading programmer stop >>>>>searching for higher NPS, better searching techniques etc? When somebody will >>>>>REALLY tackle the 2 problems I mentioned? Because otherwise a computer can still >>>>>be beaten in 2010, running on 500 GHz. But as I already mentioned: This is the >>>>>computerchess paradox: NOBODY wants to sac NPS for more knowledge. And as long >>>>>as nobody wants to quit this 'rule', human GM's are still superior in knowledge >>>>>and understanding of the game. >>>>> >>>>>Jeroen >>>>> >>>> >>>>The speed vs. knowledge dilemma is a false one. >>> >>>Wow.. now that's a statement. >>> >>>>It may apply to Rebel and other programs, but it doesn't apply to >>>>Junior, where I have a framework to code evaluation stuff virtually >>>>for free. >>> >>>Let me guess, pre-processing...? >>> >>>Ed >>> >> >>Junior roots are in preprocessing, and it was a preprocessor until '94. I do >>something better now. >> >>With your huge experience, do you seriously believe that a program on Junior's >>level can be a preprocessor ? > >I believe in theory it is possible to have good results with a preprocessor >system that incrementally updates its values according to the position in the >tree. It's known from Firtz5 and below that it doesn't evaluate positions but >moves which makes it extremely fast. It's known from Genius it has massive >pawn structure stuff at the root all preprocessor related. So I think the >answer to your question is yes. > >Not that I prefer such a system but that is another case. > >Ed > >>Amir
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.