Author: Robert Hyatt
Date: 05:47:33 09/15/04
Go up one level in this thread
On September 15, 2004 at 00:31:02, David Dahlem wrote: >On September 14, 2004 at 21:57:39, Robert Hyatt wrote: > >>On September 14, 2004 at 17:39:28, David Dahlem wrote: >> >>>On September 14, 2004 at 17:10:39, Robert Hyatt wrote: >>> >>>>On September 14, 2004 at 11:32:48, David Dahlem wrote: >>>> >>>>>On September 14, 2004 at 10:30:07, Robert Hyatt wrote: >>>>> >>>>>>On September 14, 2004 at 01:46:33, martin fierz wrote: >>>>>> >>>>>>>On September 13, 2004 at 10:20:05, Robert Hyatt wrote: >>>>>>> >>>>>>>>On September 13, 2004 at 06:41:10, Tord Romstad wrote: >>>>>>>> >>>>>>>>>On September 12, 2004 at 03:52:56, Sune Fischer wrote: >>>>>>>>> >>>>>>>>>>On September 12, 2004 at 00:42:50, Robert Hyatt wrote: >>>>>>>>>> >>>>>>>>>>>Sure, forcing Kasparov to play English openings, forcing Karpov to play the >>>>>>>>>>>Latvian, etc. Wouldn't be very revealing however... >>>>>>>>>> >>>>>>>>>>Indeed it would, every player should know his strong and weak points, perhaps he >>>>>>>>>>plays the english better than he is aware of :) >>>>>>>>> >>>>>>>>>At least he (Kasparov) apparently plays the English better than you and Bob >>>>>>>>>are aware of. ;-) >>>>>>>>> >>>>>>>>>The English opening was one of his main weapons with white around >>>>>>>>>1985-1990. >>>>>>>>> >>>>>>>>>Tord >>>>>>>> >>>>>>>> >>>>>>>>The point is he doesn't choose to play it a lot today. Nor other openings. >>>>>>>>Remember that he claimed that was the thing that cost him the last game against >>>>>>>>DB, playing an opening he didn't play much. >>>>>>> >>>>>>>i don't understand your view on the whole subject; but i specially disagree with >>>>>>>your statements about chess (players). modern grandmasters have an incredibly >>>>>>>broad opening repertoire. this is a rather new development, probably assisted by >>>>>>>chessbase. if you believe that korchnoi never plays anything but 1.d4 or that >>>>>>>kasparov never plays the english in serious games, you are rather mistaken. >>>>>> >>>>>>I didn't say "never". I said "in important games". That is pretty true. I >>>>>>have asked this very question to three different GM players, all said that they >>>>>>have specific favorite openings for important games/matches. And all said they >>>>>>do _not_ play every opening system, which means your "incredibly broad >>>>>>repertoire" might be true in general as they do study openings a lot, but it >>>>>>doesn't appear to be true when it is time for optimal results as in important >>>>>>matches or tournaments... >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>>kasparov was looking for excuses everywhere to explain his failure in the DB >>>>>>>match, it sounds more like a plausible attempt of his to cover up the fact that >>>>>>>he just played a horrible match. >>>>>>> >>>>>>>here's why i don't understand your attitude in this thread: as a chess player, i >>>>>>>have learned a lot by playing different openings. i have, for example, lost my >>>>>>>fear of the IQP over the last years as a result of playing the tarrasch defence. >>>>>>>now the tarrasch is supposed to be just very slightly dubious, but it doesn't >>>>>>>matter: i learned something there. by restricting crafty to what you know it can >>>>>>>do you might be missing things it can't do that you might otherwise find. >>>>>> >>>>>>All that is well and good. But the point is still this: Time is limited. And >>>>>>one way to fix a problem is to bypass it. IE in building a road, I can build a >>>>>>bridge over a lake in the way, or I can build _around_ the lake. The latter is >>>>>>way faster and cheaper, and the net result is the same - you still get to where >>>>>>you want to go. >>>>>> >>>>>>I've said many times in the past that I _have_ done what you suggest. You might >>>>>>remember my comments about Cray Blitz _never_ playing a g3/g6 type opening as it >>>>>>didn't understand the importance of the bishop in those positions. It was >>>>>>easier at the time to just not play them. I later fixed that in Crafty and see >>>>>>no problems with that kind of play now. But I see no overwhelming need to fix >>>>>>every kind of opening setup immediately, when it can be deferred until some >>>>>>later time with a simple book fix. >>>>>> >>>>>>That is my point. Not that a program _shouldn't_ play all openings well, but >>>>>>that there is no rush to make that happen in every case. >>>>> >>>>>So why should engine testers avoid the openings that an engine doesn't play >>>>>well? To do that is not an accurate method of testing engine strength. To get an >>>>>accurate picture of engine strength, testing must be done on a wide variety of >>>>>opening lines, not just a few lines the engine author likes. >>>> >>>>Again, why don't we settle this under "equal" conditions on a Karate mat? We >>>>get the same equipment, clothing, etc. >>>> >>>>Or does "experience" factor into that. But not into chess??? >>> >>>This is not even close to being responsive to my post. Instead you change the >>>subject to ??? Karate ??? :-) >>> >>>Regards >>>Dave >>> >> >>I think "most" get it. I challenged you to a contest that is equal by any >>external measure you can name. Equal equipment. Equal everything. _except_ >>for the "engine". Surely such an equal contest will be fair? >> >>The point is that if you make a program play oddball openings, it can/may have >>trouble because of decisions the author made during the development. Can't >>handle fianchetto openings? Exclude them from the book for the time being, and >>by doing so the engine plays _much_ stronger. Making it play them artificially >>lowers its apparent strength relative to other programs, even though in _real_ >>contests it will be as good or better because of the opening selection. >> >>That was and is my _only_ point. Playing _MY_ program with oddball books >>doesn't really show how well or poorly it plays. Because it is forced into >>unexpected territory too often... > >You are calling standard eco opening lines "oddball". They are only oddball to >you. 1. g4 is in ECO. 1. f3 2. Kf2 is in ECO. They are _definitely_ oddball. But the definition I have used for oddball is, and always has been, "any opening a program doesn't like playing naturally." This might include the English opening as white which can get very blocked, or the Najdorf lines where a single mistake loses the game, or whatever... I have been _completely_ consistent in that definition and usage, from day 1. If an engine doesn't know how to play some standard eco opening lines, that >is a weakness of the engine, and this weakness should be taken into >consideration when evaluating the overall strength of the engine. Case Closed!! or "mind closed" perhaps? Several including one commercial book author have pointed out the problems. If you choose to ignore them, that's fine. But ignoring them/me does _not_ make "case closed" by any stretch... > >Dave > > >> >>I limit that with _my_ opening books. Sandro does this for Shredder. Alex for >>Fritz. Jeroen for Rebel. Etc. Otherwise why does everyone need a person >>responsible for the book if it makes any sense to rip that book out and force >>the program to play with something that is not tailored to its specific playing >>style??? >> >> >> >>>> >>>> >>>>> >>>>>Regards >>>>>Dave >>>>> >>>>> >>>>>Why is that so hard to >>>>>>understand? _everybody_ has done it and still does. That is what the >>>>>>professional book authors contribute to a chess development project... >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>>besides: how many players are interested in a strong chess engine? on my level >>>>>>>(FM) i know exactly ZERO people who play against engines. on the other hand, 99% >>>>>>>of the poeple i know on my level use engines to analyze games. conclusion: if >>>>>>>you want to make an engine that is useful for the average user of strong chess >>>>>>>engines, it has to be able to deal with all positions that the user throws at >>>>>>>it. >>>>>> >>>>>>supply and demand. How many FM/IM players are there vs how many club players? >>>>>>It is the market that drives this. Chessmaster is the best example. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> >>>>>>>cheers >>>>>>> martin
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.