Author: Dann Corbit
Date: 11:33:39 05/02/01
Go up one level in this thread
On May 02, 2001 at 14:21:18, Chris Carson wrote: >On May 02, 2001 at 14:09:53, Dann Corbit wrote: > >>On May 02, 2001 at 13:48:04, Chris Carson wrote: >> >>>I would like to hear from the CCC group how much K will >>>gain from having the program for 3 months. In my view, >>>an advantage yes, but maybe not as much as I thought at >>>first. >>> >>>1. K may not have the HW for the match. >>>2. To get a real feel for Fritz 7 he will need the HW and >>> play 40/2 games. >>>3. K has to prepare for other opponents for matches and >>> tournaments, I think this will be a higher priority for K. >>>4. K has other commitments than just matches and tournaments that >>> will also be a higher priority. >>>5. K may gain a sense of how Fritz should play at any given time, >>> and may be over confident by match time. >>> >>>None of the above are guaranteed, but in my opinion, lots of games at >>>fast time controls is not the same as 40/2 on the real HW. Also, my >>>guess is that K will play some each week against what ever version of >>>Fritz he gets, but each will be slightly different (ala play different >>>versions of crafty with slightly different books) and not the same. >>> >>>Any thoughts. >>> >>>I do think this gives K a huge advantage, I am just not sure he will >>>have the time to fully utilize it. Just some thoughts. ;) >>> >>>My guess now: K wins 6.5-3.5 with a couple of surprises by both >>>and a couple of blunders by both. ;) >> >>I don't think he would just try random openings or something silly like that. >>He will aim for strategies. In other words: >> >>1. Will it fall to a giant material imbalance? >>2. Will it fall to blocked pawn formations? Can these formations be easily >>created and if so, in what manner? >>3. Will it have a NULL move blindside that is repeatable? >>4. Are there early positional moves it will not understand? >> >>That sort of stuff. Armed with data like that, I don't think you could give >>them a larger advantage. I don't think GM's have worked very hard at beating >>computers. What I mean is, they try to beat them conventionally. I suspect >>that if they carefully studied Raphael Vasques' Anticomputer site, their ELO >>against the machines would instantly go up 100 points. And if they worked on >>anticomputer strategies it would go up another 100 points. And if they spend >>three months preparing for a specific opponent, it will go up another 100 >>points. You might imagine that they already do this stuff. However, I have had >>email conversations with GM's after they engaged in high-profile matches against >>computers. I was surprised to learn that some are actually unaware of the value >>of closed formations and some of the basic anticomputer strategies. > >I agree with you Dann, that represents the upper bound for Kramnik and >doom for Fritz. It may come to pass if Kramnik fully takes advantage >of everything and the Fritz programmers have no surprises for him. > >I am not convinced that Kramnik will devote as much time preparing for >the computer (as you pointed out that other GM's have not) that is afforded >to him and I also think the Fritz programmers may have a wrinkle or two >for him. This will make no difference in the outcome. Kramnik will win >big, but it may not be a blow out like 10-0 or 9-1 or 8-2. Kramnik starts >with about 100 pt ratings advantage. Give 300 more as you have pointed >out and that is 400 points (maybe more). Thus you would expect 9-1 or >better for Kramnik. 2 or more points for Fritz would have to be considered >an accomplishment. Agree or Disagree? :) I agree with you. Unfortunately, I'm wrong a lot! ;-) That is one reason I find the contest so interesting. There may be some clouds over the whole thing, but I will still be drawn to it like a moth to a flame. I lack the moral strength to boycott. I expect wonderful chess and wonderful surprises in any case -- blowout or no blowout.
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.