Author: Christopher R. Dorr
Date: 12:38:12 09/07/01
Go up one level in this thread
On September 07, 2001 at 15:17:20, Roy Eassa wrote: >On September 07, 2001 at 09:43:00, Christopher R. Dorr wrote: > >>I really do not see these conditions as a huge advantage for Kramnik. >> >>1. If Kramnik was going to play a serious match against a human GM, he would >>study all the games by him/her in depth. Most human GMs would have hundreds or >>thousands of games to study. Deep Fritz does NOT have these games yet. A >>reasonable alternative would be to allow the GM to use the program before hand. >>The CB team certainly has studied *all* of Kramnik's games closely. Why should >>the computer have the advantage of knowing the GMs style and weaknesses when the >>GM does not get to understand the computers? >> > >The GM is not just going to UNDERSTAND the computer, he is able to prepare >dozens of EXACT WINNING GAMES in advance, the moves of which Fritz will repeat >in the match. The computer cannot "adapt" to this strategy. (Even if his >personal copy of Fritz has "learned" to avoid certain openings, that learning >will not carry over to the copy used in the match.) So Kramnik has an advantage because of the inherent construction of a semi-deterministic machine? The CB team won't be able to modify the book so that the opening probability weights are modified in the match book? To get to the same middlegame, the same opening moves must be played. Somehow, Kramnik is going to remember multiple games in *every* variation of the Sicilian the box might play? Every subvariation? Every variation of the French? Every Subvariation? Every variation of the Caro? Every subvariation? Every variation of the Spanish? Every subvariation? And he's going to play them at the exact same time that he did in his preparation, so that DF can't go 1 ply deeper and see something else, or ends up one ply shallower, and plays something different? Impressive. > >>2. Fritz can change it's book at will. If there *are* any 'predetermined' games, >>they arise from the opening. Modify the book, elimninate the problem. >> > >Nope. There are always going to be oddball moves that take the computer out of >book immediately. Remember, Kramnik does NOT need to figure these out during >the match. He has MONTHS to work out numerous odd lines IN ADVANCE that he >KNOWS the compuer will mis-play. Once again, the computer cannot "adapt" to >this technique as a human might. As I said in another post, you can certainly throw in some trash openings. If Kramnik plays 1.h4 in his prep and comes out with a winning line after the machine plays 1.Nf6, what happens when the DF people toss in a few lines where after 1.h4, DF plays c5? Suddenly, Kramnik has played a terrible move, hoping for a trap that now won't come. Who has the advantage now? This isn't some 3 0 ICC match. There's a million bucks on the line. Kramnik's going to risk that kind of nonsense, when he has the built-in advantage of being a great positional player? > >> >>3. And I doubt there could be predetermined games anyway. Think about the >>variety of openings. Kramnik is going to *find* and *memorize* killer lines in >>everything? As black, he would have to prepare for at least 12 major openings by >>white, and literally hundreds of important subvariations (If he chooses to play >>a Scheveningen Sicilian, for example, you need to worry about Keres attack with >>h3, Keres Attack with Rg1, Classical, Bc4, King's Indian Attack......). And >>remember, there are already very few *big* holes in Fritz's opening book. The >>*small* ones Kramnik would find would likely be further down the tree, making it >>even more unlikely that he could spring a particular trap. >> > >See last response. Kramnik could avoid "good" openings altogether and play >utterly "bad," unknown lines that a strong human might refute positionally; the >computer's responses can be KNOWN IN ADVANCE and Kramnik and his team have >plenty of time to find REPEATABLE variations in which Fritz is KNOWN to play in >a losing way. Even if only 1 in 50 "bad" openings work against Fritz, months is >plenty of time to find and learn those 1 in 50. [Somebody reputed to be Bobby >Fischer has been annihilating IMs on ICC using openings that are horribly bad, >like f3, Kf2, Ke3 as White's first 3 moves. If it's possible to recover from >such horrendous moves against a strong human (admittedly in a fast game), it's >certainly possible to find less-horrible lines that will never be in any >computer's book yet that the computer repeatedly mis-plays.] > And again, as soon as the DF people have modified the opening book with some new responses to trash, then Kramnik is playing 1.f3 against a FIDE 2600. Think he'll risk that with a million dollars on the line? That would be insane. >> >>Overall, I do not see Kramnik as getting a huge advantage. He didn't get the >>source code. The program isn't under a 'truth serum'. He has access to it's >>play, just as CB has access to all of Kramnik's games. As matches go, I see this >>one as fairly equal in it's treatment of the two players. > > >The source code would be nearly useless to Kramnik. But for the reasons I >mentioned above, the match is hardly fair or even. > > >Note: Kramnik MAY very well choose to take the high road and avoid the above >methods. I even think it's more than 50% likely he will. However, I am quite >certain that, if he and his team of helpers wanted, they could win EVERY GAME >against Fritz by the above methods. Doubtful. Again, modification of the opening book makes trash openings useless. Playing against style against a 2600 computer will get one killed fast. Want to know how Kramnik will win this match? Same way he would have won this match without ever seeing a copy of DF. By playing good, sound positional chess (like that which helped him become world champion) and outplaying an opponent who is almost certainly weaker than he. It's that simple. Chris
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.