Author: Uri Blass
Date: 11:05:22 08/24/01
Go up one level in this thread
On August 24, 2001 at 11:55:15, Vincent Diepeveen wrote: >On August 23, 2001 at 14:36:16, Dann Corbit wrote: > >>On August 23, 2001 at 14:26:01, Gian-Carlo Pascutto wrote: >> >>>I got the following anonymous message in my mailbox >>>from another amateur: >>> >>>>>> >>>dat van vincent over openingsboeken op CCC gaat idd over deepfritz.ctg, goliath >>>gebruikte dat namelijk ook tijdens Leiden '01. Die had dus keurig deepfritz.ctg >>>in zijn interface staan, maar dat verborg ie telkens nadat ie de interface >>>opstarte >>>>>> >>> >>>Bah! This is the kind of crap why I don't enjoy computerchess >>>anymore. >> >>Use of commercial books is old hat. You will find that quite a few amateur >>programs use the fritz 4 book, and have done so openly for many years. >> >>If it is not spelled out as a rules violation, then it is not wrong to do so. > >Let's be clear here > a) if someone is lying then that pisses me off (as the truth sometimes > is so simplistic to see) > b) if someone WINS because of a certain book which he simply > only cut'n pasted then that pisses me even more off. > >If someone uses either > shredder5, fritz4, genius5 or whatever similar book, then i won't complain > soon, because those books aren't so called 'killerbooks' or already > outdated by now. > >Starting with Alexander Kure (fritz6, deepfritz and nimzo books) a new >generation of books came onto the market which simply surgically annihilates >and refutes all lines from automatical generated books. > >Note Noomen also does a great job here, but the real difference >between Noomen and Kure is that the Noomen book isn't put onto a CD >which everyone can buy. > >Of course the Kure book only works with the chessbase GUI, so things >are pretty easy. > >Writing over a few lines by hand is a different thing. this is labour >intensive, and hard to prevent. The real bad thing is that from those >5.5 points from Gromit at least 4 of them are because of Kure, because >it is not bad in endgame (and positions where material has gone off), >but without Kure it would never have gotten into such positions. > >Note i don't want to complain too much about Gromit. Best thing is >to prevent this thing from happening in the future. > > a) otherwise most people who make their own book feel like they > are busy doing something useless > b) It is simply unfair > >>It probably *should* be specified one way or the other. I doubt that it has >>been or the problem would not arise. > >It has to do with general stupidity too and the fact that outlaws >count on that people who try to keep the law are in this case professors >who are dead lazy and who know themselves shit from books or whatever >happens in computerchess, but especially the outlaws count on that >walking on the edges from what is legal and considered not fair is >not actively enforced by the tournament director. > >Note in this case the tournament director did get a protest from me, >but as i told him a general protest for the future (with no effects >on what happened in the past) is going to get shipped. > >Of course i know how germans think: "let's do it we might get a world >champ". In general the only dudes that used commercial books last >few years in world champs without clearly having asked permission to >the authors were all germans. > >I remember SOS who openly won the amateur title using a Kure book. >Note he didn't hide a thing and only after the world champs was over >he started realizing that the majority of his points were made by him >because of this. > >This year the amateur winner is no different and in i-csvn tournament >the program Goliath was also using something called 'deepfritz.ctg' though >clearly some lines were modified (as behind the deepfritz.ctg book he >also used a hand made book). > >This year SOS used the shredder5 book and started with 0.5 out of 4 >at the tournament. > >Also amazing was the draw out of book from Isichess versus SOS. > >The only engines from which i can remember in the worldchamps that they >ran under a different dudes interface: > - Gromitchess (deepfritz) > - Goliath (deepfritz) > - SOS (shredder) > >the first one i suspected using deepfritz.ctg and a few useless >additions. > >The second is always suspected but was warned after i-csvn tournament > >SOS openly used shredder5 book, but joined as a semi-professional and >is getting sold together with shredder5 and definitely was using >a that bad book (necchi + automatic generated crap) that no one >would care. in fact most are very happy if you use a shredder5 book >as it's complete suicide to use an automatic generated book and Necchi >should be fired by Stefan Meyer Kahlen anyway. > >What do i need to say more about this point, 3 out of 3 is a 100% score? > >>Let's congratulate the winners, and put a bit of effort into improving the >>system. Perhaps we can send email to Beal or whoever is in charge respecting >>requested changes for the future. > >v/d Herik in this case. > >>At any rate, an entertaining tournament as always. Shredder has once again >>risen to the top, and certainly deserves every accolade. > >>Deep Junior was very impressive also. > >Junior7 plays like major shit. I'm amazed that they won the world title >by such a kind of swindling play. In fact it's quite funny they won it, >as it shows how much you can reach by a bit of consequent swindling. > >8 out of 9 is a major score! Junior7 seems to have some knowledge that other programs do not have. Junior7 had a good game against Crafty when it had better evaluation than Crafty. Junior evaluated the attack chances as better than the pawn advantage of Crafty and the result suggest that Junior was right. Junior has also a bad position against Shredder and won the game when your program got a better endgame against shredder and could not find g4. g4 wins and Deep Fritz can see a clearly winning score after some hours(I remember more than +5 at depth 24) I do not understand what your program saw at that position. Deep fritz always fail low if I give it the position after g4. It does not solve the fail low and the score droppes by 0.31 or 0.32 every iteration. I do not believe that your program could see in some minutes more than other programs see in some hours after g4. > >>And Gromit has obviously made enormous strides. > >??? > >I only remember they started progressing when they ran under >chessbase environment. > >If they wouldn't have had a completely strategically winning position >against me (from computer perspective, though objectively it >was ok for white), i would have completely annihilated them! > >Idem for many others. > >We do not know how they play if they have to play on their own! > >We only know that SOS who played with a weak book this year started >with 0.5 out of 4. I think that it is better to improve your engine instead of crying about books. If your engine is good enough then a small book that was generated manually by you is enough to win because the engine is going to find the good moves when it is out of book. It is also possible to get programs out of book in few moves by lines like 1.e4 e5 2.Be2 and I expect programs to get 50% If your engine is weaker without book than other programs than the problem is with the engine. Uri
This page took 0.01 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.