Author: Sandro Necchi
Date: 06:37:43 01/09/05
Go up one level in this thread
On January 09, 2005 at 08:58:41, chandler yergin wrote: >On January 09, 2005 at 03:40:06, Sandro Necchi wrote: > >>On January 08, 2005 at 22:35:28, chandler yergin wrote: >> >>Hi, >> >>>On January 08, 2005 at 21:40:12, odell jones wrote: >>> >>>>On January 08, 2005 at 20:21:56, chandler yergin wrote: >>>> >>>>>On January 08, 2005 at 18:00:30, Derek Paquette wrote: >>>>> >>>>>>We've seen Junior and fritz play kramnik and kasparov,.. fritz more than once. >>>>>>Does anyone know if Kasparov/kramnik/topolov will ever get to play Shredder? >> >>I agree with you that ChessBase should chose a different program than Fritz to >>play Kasparov as we saw Fritz more than once, but maybe "someone" in ChessBase >>is a fun of Fritz...:-) >> >>Chess Base should select the program that is the strongest to face the strongest >>chess player. >> >>Since Shredder is leading the SSDF list since years now, we do not see articles >>on Chess Base site on this anymore...maybe it is not news anymore:-) >> >>>>>> >>>>>>At the Men vs Machine world championship, shredder was absent. >> >>Even if asked we would have refused as we had to work on the new Shredder 9. >> >>>>>> >>>>>>when are we going to get to see shredder play a MATCH vs a super grandmaster, >>>>>>and a high profile one at that. >> >>I hope soon! >>It is only a hope as, as far as I know, there is nothing planned on this yet. >> >>>>>> >>>>>>Anyone heard anything? Anyone know anything about this subject?? >>>>>> >>>>>>Why not have the strongest computer program playing the humans, atleast ONCE. >> >>I agree 100%!! >> >>>>> >>>>> >>>>>The ELO Performance Rating of the TOP Programs approaches ELO = 2600 >> >>This is not correct. >> >>Not considering Hydra, which has a very powerful hardware and talking about the >>top commercial programs the Elo is the following: > >Where would Hydra fit into this Category: > >Hiarchs 8 Opening Tree based on 1,323,335 Games Ave Elo 2403 Perf. 2436 > >Fritz 8 Opening Tree Based on 122,921 Games Ave Elo 2573 Perf. 2619 > >Fritz 7 Opening Tree Based on 239,391 Games Ave Elo 2549 Perf. 2582 > >Shredder 8 Opening Tree Based on 520,335 Games Ave ELO 2510 Perf. 2544 > >Fritz 5.32 Opening Tree Based on 110,503 Games Ave Elo 2537 Perf. 2584 Nobody knows as the data is not available and secret. The last book may have something like 200,000 book moves as it is limited to 10/12 moves (20-24 plies). > >Who has determined Hydra has a Performance Rating of over 2800? I do not know. I believe it is less; about 2700 against humans and 2800 against computers (My opinion). > >Who has determined Hydra or any other Program has an Elo of 3000? Nobody said that. I said that depending on how well the opening and resulting middle games and endgames the performance (strenght of play) can change from 2200 to 3000. This is based on tests made by me. > >How can a Program have an Elo of 200 Points Higher, than the Opponents >it plays? I am talking about performance...some players do perform much higher on some tournaments than their Elo... See Kasparov, just to name one in the latest Russia Championship... > > > > > >> >>2200 if the opening is very bad and the position is very bad for them; they do >>not know how to handle it. >> >>3000 if the opening is very good and the resulting middle-game/endgame is >>suiting the program. Also using a powerful hardware. >> >> >>So, the trik is to bring the program in the best suiting positions as much as >>possible to get at least 2800 Elo. >> >>>> >>>> >>>> >>>> >>>> >>>>Where do you get 2600 from??? The Evidence suggest much higher >>> >>> >>>I refer you to a Post and Dialog I had on CTF >>> >>> >>>Thanks for input.. >>>>>Hiarcs 8 appears to have the largest & up to date Opening Book of all the Top >>>>>Programs. >>>> >>>>I disagree on that, the size of an opening book has very little to do with >>>>quality. Currently there are 3 opening book experts in the world, 1) Sandro >>>>Necchi (Shredder), 2) Aleander Kure (ChessBase), Jeroen Noomen (Rebel, Chess >>>>Tiger). The author of the Hiarcs opening books I don't consider as a leading >>>>expert. >> >>The size of the book has nothing to do with the quality of the book. Quality >>means selection and "changes" as in the current theory there are a lot fo >>mistakes (moves/variations which are very weak and are shown good). >>If we take the openings books of top programs (Shredder, Fritz, Junior & Chess >>Tiger; just to mention the most famous ones) we can see that there are a lot of >>moves which are not in "human theory" or that the given evaluation is different. >>This means that the book authors have been working on them. Some of them since >>years... >> >>Also when we all meet at the WCCC you can see that the openings books are >>different than the commercial ones and so it would be the same when facing a >>strong human player. >>This is forced to avoid some weaker lines which are in the commercial books to >>increase variety and not to let the opponents know what we would play to avoid >>the opponent prepare against us. >> >>Kramnik asked to get the program in advance and the book also before the match >>against Shredder some years ago. >>We refused. >>This request was changed facing Fritz. >> >>>> >>>>I am not sure about GM Boris Alterman (Junior) and Dan Wulff (Gandalf) as true >>>>experts, they could be but I don't know their work quite well. By a true expert >>>>I mean those people who are walking encyclopedia's when you talk with them about >>>>computer opening theory. Or, when you hear them debating, you don't believe your >>>>ears, they know every dusty place of their large opening books of often more >>>>than 500,000 positions. >> >>Only 500,000 positions? >> >>>> >>>> >>>>>Yet Shredder 8 keeps on going and winning.. so it has to be something >>>>>else. >>>>>What? >>>> >>>>Better search + better chess knowledge. What else? >>>> >>>>Ed >>> >>> That must be it! Congrats to the Programmers! Awesome! >>> >>>I was looking for a relationship between the Opening Tree & Performance. >>> >>>Perhaps you can comment on this.. or see some relationship >>> >>>Hiarchs 8 Opening Tree based on 1,323,335 Games Ave Elo 2403 Perf. 2436 >>> >>>Fritz 8 Opening Tree Based on 122,921 Games Ave Elo 2573 Perf. 2619 >>> >>>Fritz 7 Opening Tree Based on 239,391 Games Ave Elo 2549 Perf. 2582 >>> >>>Shredder 8 Opening Tree Based on 520,335 Games Ave ELO 2510 Perf. 2544 >>> >>>Fritz 5.32 Opening Tree Based on 110,503 Games Ave Elo 2537 Perf. 2584 >>> >>> I note also that the Source of the Fritz Games are those >>>published in Chessbase Magazine. >>> >>>Each Version of Fritz of course has more Chessbase Magazine published games. >>> >>>Anyway... I'll turn 'my' Ponder off, and wait for your comment if any. >>>;) >>>Chan >>> >>> So, Odell, If you can "refute" the above..please do so.. >>>Or >>>STFU! >>>CY >>> >>> >>>>> >>>>>Kasparov is/was = 2830 The strongest Player in History! >>>>> >>>>>Progerams have NO intelligence.. they Store & Retrieve Data! >>>>> >>>>>What they play in the Opening is GM Moves.. played before! >> >>This is not true at least since 8 years. I am talking about the top chess >>programs and not only about Shredder. >> >>The reason most people do not know this is because they mainly test without >>books or with books made from games databases... >> >>>>> >>>>>Why would those who know this.. have much interest in a Match like that? >> >>You do not know. This was the case in 1987 when I showed the GMaster opening >>book in Rome at the 7th WCCC which did include new moves and not only those >>which are in chess books and ECO. >>In fact if you take a game won by a strong GM against another one you will not >>be able to play it anymore as the GM did study that game and already decited >>what to change to improve it. >>So the only way to surprise them is to bring something new they are not prepared >>to face...I know it is not easy...but it is the only way. >> >>>>> >>>>>People go to Car Races to see someone Crash into a Wall. >>>>>People watch Human vs Computers to see the Human fall into a Position where the >>>>>EGTB's take over. >> >>This is very unlikely as they take over too late in the game... >> >>>>> >>>>>Fritz Programs have a BUG! >>>>>Other Programs have BUGS! >>>>>The Human.. using the Programs with BUGS; Figure out 'what' the BUG IS, >>>>>and learn how to play to exploit the weakness. >> >>It is not that easy...and we also know our weaknesses and work on them to >>eliminate/improve them in many ways... >> >>>>> >>>>>Cars go faster than Horses... so what? >>>>>Peple still like to watch Horses. >>>>> >>>>>Others don't care, and like to watch Cars. >>>>>So What? >>>>> >>>>>Are you getting my point? >> >>No, I think you are not updated with currently top chess programs. >> >>Sandro
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.