Author: Sune Fischer
Date: 08:11:08 04/06/02
Go up one level in this thread
On April 06, 2002 at 10:31:06, Otello Gnaramori wrote: >On April 06, 2002 at 05:03:29, Sune Fischer wrote: > > >>I am one of those ignorant people that have some doubts about the true strength >>of the programs ;) >>Let me explain why... >>We all agree that in tactics nothing beats the computers, but how good are they >>at the positional level? > >Not to be boring, but I would like to stress that tactics is way more important >in chess than positional knowledge for winning purposes I mean... Maybe it is an irrelevant example, but my little engine is now seeing two plies deeper than the previous versions. I have been playing matches against TSCP and Ozwald043, it has a good positive score, so in engine-engine matches is now clearly stronger. However I have done nothing about the evaluation, and all you have to do, is to wait for it to castle, sacrifice a few pawns so you have a clear shot at its king, and you can just roll it over! Point is, it is not playing any stronger against humans as far as I can tell, humans just go for the king, and it doesn't begin to defend itself until it's within its horizon, which is often too late. The top programs probably have this snag fixed, but there might be other similar weaknesses that one can find. It takes some time and elo to find these, but I bet there are openings/middelgames/endings that computers in general play weaker than others, it is a matter of "research". >>I am sure the programs are way behind GMs in the static analysis of a position, >>because a programmer can only implement a bunch of static rules, > >That bunch of rules is often fruit of YEARS of tuning with many, many parameters >, but nevertheless I consider much more important the overall including the >innovative search algorithms. I think the tuning is mostly done in engine-engine matches because of the amount of data needed, humans could probably also teach the thing a lesson or two :) >>and not the >>extensive knowledge and experience of a GM, often the programmer is not a very >>strong player himself, which doesn't exactly help either. > >That's your opinion, please read here: > >Martin: 2. Must a chess programmer also be a good chess player in order to write >a top program? How important is chess knowledge? > >Christophe Theron: I believe that being a strong chess player is a serious >HANDICAP if you want to write a strong chess program. If you have too much >prejudices about chess, then you are not really in the right state of mind. To >write a strong chess program you must have a very open mind and be prepared to >react in a flexible way. You must be prepared to try as many ideas as possible. >You will find that most ideas that sound obvious for a human chess player do not >work. You must be ready to give them up and to think differently. As time goes >by, you develop a different understanding of the basic nature of chess. > >Of course, chess knowledge is valuable, but the problem is that a computer >program does not need the same knowledge than a human player. You need to filter >what chess knowledge is going to be useful, and what knowledge cannot be used >(yet). As a programmer, after looking at many games played by my program, I have >been able to develop some sort of strategic sense of the game. It is nowhere >near what a GM can do, but this knowledge, implemented in my program and >enhanced by the speed of a computer, gives the strength of the Tiger engines. > >There are a number of concepts that human players use that I have found to be >useless for my chess program. The most interesting one is the concept of tempo. >It is a key concept in human chess, but it appears nowhere in the source code of >Chess Tiger. I guess that the program is able to SYNTHESIZE this concept from >the other concepts it knows, which is a very fascinating thought. Obviously I my word means nothing compared to Christophe's, but I doubt it would hurt the development of Tiger if Chris was also a GM. He says he has learned a lot of positional concepts in the process, I'm sure extra knowledge would not hurt but in fact help. >>This is a weakness in the programs, it is there but seems to be completely >>unexploited by most of the strong players. >>Exactly how strong does one needs to be to dissect a program completely is not >>clear, but I have heard of players with sub 2000 elo that has a great score >>against the programs. > >That is a legend... It is probably not as easy as that, but have we ever tried? >I think it can be done, GMs don't bother however. In all >>the matches GM vs computer we have seen lately, it's been 4-8 games, far too few >>for a complete dissection. >> >>To my knowledge programs today are mainly used for training, testing new >>openings and analysing games, that sort of thing. >>GMs are so used to playing humans, that they can't change their style when >>facing computers. > >Why should have to force themselves to play in a different way at any >cost...don't understand you. You should attack at the weakest point, is it best for computers to have queens on the board or not, best to play closed or open positions, best with bishops or knights, which phase of the game do they play the stronges, when do you complicate, when do you simplify etc etc. These questions could be answered by research, and would give GMs an edge IMO. >>Much time and energy is wasted by "thinking the wrong way". > >???? Read Gulko's comments on the match. >>Remember that 99.9% of all the games they play are against humans, this is how >>they make their living, so bad habbits die hard I guess ;) >> > >C'mon Undskyld? >>I can be convinced that programs are *truly* above 2500, but I would have to see >>the programs enter hard and serious tournaments, where players would bother to >>find the weaknesses of their opponents. > >I guess that many masters already own in their personal computer a strong chess >program where to practice...so that's excuses > >>A 24 game match against a 2400 player, lots of money at stake, with one days >>rest between games and loads of strong GM analysis of the games during the >>match, I think the IM would win. >> >>For all practical purposes; playing random opponents on the internet and very >>short matches, I must agree that good programs today perform quite consistently >>above 2500. >> >>-S. > >I 's say 2600 and more. 2600 is also above 2500 last I checked ;) >w.b.r. >Otello -S.
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.