Author: Vincent Diepeveen
Date: 11:14:37 04/24/01
Go up one level in this thread
On April 24, 2001 at 13:42:43, Robert Hyatt wrote: >On April 24, 2001 at 11:34:59, Vincent Diepeveen wrote: > >>On April 24, 2001 at 11:23:37, Robert Hyatt wrote: >> >>>On April 24, 2001 at 10:19:21, Vincent Diepeveen wrote: >>> >>>>On April 24, 2001 at 09:34:18, Terry McCracken wrote: >>>> >>>>>On April 24, 2001 at 08:20:57, Vincent Diepeveen wrote: >>>>> >>>>>>On April 24, 2001 at 03:47:15, Uri Blass wrote: >>>>>> >>>>>>>the best software that is not IBM. >>>>>>> >>>>>>>Suppose there is a match of 20 games at tournament time control >>>>>>> >>>>>>>I am interested to know how many people expect 20-0 for IBM >>>>>>>How many people expect 19.5-.5?.... >>>>>> >>>>>>>If IBM expect to do better result then the average result that the public expect >>>>>>>then they can earn something from playing a match of 20 games with Deep Blue. >>>>>>> >>>>>>>I believe that a part of the public who read the claim that kasparov played like >>>>>>>an IM are not going to expect good result for IBM.> >>>>>>>Uri >>>>>> >>>>>>First of all IBM would get out of book every game with -1.0 pawn >>>>>>disadvantage (which is about the average of what Kure and Noomen >>>>>>get in tournaments, sometimes they get out of book with mate in XXX even). >>>>>> >>>>>>I would expect IBM to lose with 18-2. >>>>>> >>>>>>Let's be realistic >>>>>> >>>>>> a) IBM searched 11-13 ply in 97, nowadays programs search deeper >>>>>> b) their book is hell worse as nowadays books are >>>>>> c) positionally it never was good, it doesn't even >>>>>> know what a good bishop is nor when a doubled pawn is >>>>>> good (f2,g2,g3 pattern happened twice in games against kasparov) >>>>>> also it exchanges sometimes queens in a position where not exchanging >>>>>> wins for IBM >>>>>> d) hardly can use EGTBs >>>>>> >>>>>>So in *all* respects it is getting outgunned. Not to mention EGTBs. >>>>>> >>>>>>No one talked about that subject yet, but last so many plies they can't >>>>>>use EGTBs. They only can use them the first 5 or 6 ply, that's it. >>>>>> >>>>>>Though this is very good compared to not using them, this means simply >>>>>>that all exchanges towards a lost 5 men they will not detect. >>>>>> >>>>>>They lose with induction to everything. The level of software has increased >>>>>>bigtime when compared to 1997. Of course the strategical problems are >>>>>>still there and some positional problems are still there, but in >>>>>>computer-computer games you hardly can take advantage of that. Only >>>>>>a human versus a computer can! >>>>>> >>>>>>Best regards, >>>>>>Vincent >>>>> >>>>>Hello Vincent, some of your data on IBM is faulty; >>>>> >>>>>(a)IMB's Deep Blue 1997 version could search up to 22-24 plys in the middlegame! >>>>> No commercial program can do that at an average of 3min. per move. >>>> >>>>Actually some lines diep searches up to 60 ply easily. >>>> >>>>60 ply maximum search depth is for like a 9 ply search alrady valid. >>>>Deep blue had a hardware limit of 32 ply so in that respect i always >>>>search deeper as deep blue! >>> >>> >>>DB had no such limit. I don't know where you got that from, but it wasn't >>>from the DB team or anything they wrote. They _claimed_ to search to 45-50 >>>plies along critical pathways due to singular extensions, which shoots that >>>"32 ply limit" down in flames... >> >>It might be a hardware limit not a software limit. >> >>Note i get on average always 60 ply lines at any given moment, even in >>blitz. up to 100 plies i see less. >> >>This is because i extend more as deep blue probably did. As everything >>is in software last few plies especially i trigger many extensions! >> >>>> >>>> >>>>>(b) There opening book(s) were enormous and I'd bet larger than any programm >>>>> on the market. They would not disclose how large, but maybe Joel Benjamin >>>>> could help you out there:) >>>> >>>>Yes like my 10 million move book automatically generated was. >>>>This book on average got out of book -2.0 or something against Kure / Noomen. >>> >>>However, _they_ had several GMs working on the opening book. I doubt Kure or >>>Noomen would claim to be better... >> >>Kure + Noomen definitely are better making chessprogram opening books >>as those GMs. >> >>Note most GMs play like children in lines they do not know and we >>all know how bad programs react on book. >> >>May i remind you crafty at the world champs playing Najdorf for 3 times >>and losing it 3 times chanceless, despite that in the Fritz game it >>had a few chances as fritz played some bad moves? > > >May I remind you that we had a broken book there and we could not replace >that wide book with our more selective book. And with learning disabled due >to a problem I didn't notice, yes it would play the same opening. With the >normal opening book, crafty will _never_ play the Sicilian opening as black. >And it won't play losing lines as white. But I wasn't there and we had the >problem... It doesn't matter. Crafty fell for the same opening where it can't play for 3 times! Book fix or not. that's 3 zero's in a tournament where you could easily save at least 2 zero's. >But come try it on ICC and show me how to get a winning line 3 games in a >row with any program or book you care to use... Next tournament you join Bob you'll see some! Note i'm about worlds worst book maker considering diep's achievements in opening. I just copied Nimzo's game in diep-crafty! Note i didn't even do it intentionally. I played with a wide book in world champs, but it appeared to be bigger as that of crafty. Crafty was soon out of book so it castled! >> >>Basically all those moves played by crafty there is what we could expect >>also if a GM works on its book! >> >>Junior book is also made by a GM i heart whispering in 1999. >>I do not know whether this is still sure, but i know that many lines in >>its book were not very good for it. Of course you are right in >>*claiming* that potentially a GM book *can* be better, but that's like >>me claiming that my chessprogram is infinitely better as any other >>program because i am a programmer with 2285 rating and i beat any other >>chessprogrammer with me 1 minute and he 5 minutes on the clock, >>as proven recently in IPCCC2001 :) > >That is too broad a statement. David Levy is/was a chess programmer. I'll >bet you can't beat him at 5-1 time odds. IM Mike Valvo is another. Marty >Hirsch is another. David Slate is another. This list can get quite long >in fact... Well Ossi is said to have had 2200+ in past. He lost 5 to 1. If you never play you lose your chessskills, just like you lose your condition if you do not play soccer anymore. Note also among my victims were Stefan, Franck and Matthias isn't Matthias german rating around 2000 or something? Matthias *is* an actively playing chessplayer from the above list. Positionally spoken i had most problems with Matthias as he setted up his games best of all players. Ossi won some games as he knew how to win time... David Levy doesn't play active chess nowadays so it's not fair to challenge him at this level as he will give away material too easily at that quick level. If you give away material, even if it's a pawn, you're chanceless whatever level you play of course. Basic point is not the score or whether i win 1 vs 5 easily. The basic point is that i'm a much better chess as active chess programmers. You mention all chessprogrammers who aren't developing right now a new engine themselves! Hirsch i have never seen, so i can't discuss, all i know is that mchess doesn't even run under windows, so he probably is an even worse programmer as i am! Note that mchess did positionally play way better as most other programs did at a time when mchess was still getting released. Bottom line is that letting a GM make a book is better as letting a programmer make it, that's the same statement fundamentally spoken as saying that i can make a much better chessprogram as the commercial programs which you can get in a shop now! Best Regards, Vincent > >> >>Those 2 claims are very the same. Potentially i can write a program with >>way more chessknowledge as any author who doesn't have the masterclass >>experience which i have. >> >>This doesn't mean that within 1 year of work i do create a better program! >> >>In fact i'm already busy 7 years now and only now i feel it's getting >>real strong at tournament level! >> >>Now this GM who was said to have made the deep blue book is to start with >>a very weak GM. This is not relevant. But it also was a very old GM, >>so he probably knew nothing from computers. And learning to imagine >>how a program reacts on a game is hard. > > >Joel Benjamin is _not_ a "very weak GM". His rating and tournament results >prove this.. > >> >>Also they only TESTED a few games with deep blue. I remember they played >>some blitz games against Rebel. >> >>Now that means that the book for sure was very UNTESTED! > > >Joel played against it _every day_ working on the book and the evaluation. >I don't know why you think they only tested it for a few games vs Rebel... > > >> >>I know *exactly* how an untested book performs. >> >>See the horrors that happened to me at IPCCC2001! >> >> >>Best Regards, >>Vincent >>
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.