Author: Uri Blass
Date: 00:51:15 02/08/01
Go up one level in this thread
On February 08, 2001 at 00:50:29, William Penn wrote: >On February 06, 2001 at 18:46:21, James wrote: > >>On February 06, 2001 at 15:27:23, Jorge wrote: >> >>>I have settings for CM6666, and it is very good! It is winning a tourny that i >>>made (+12 -3 =8) with different personalities, including CMKarpov, CMNimzo732, >>>CMQueen++, Bendorz, Fischer, and few others. I have CM8000 too, but I get the >>>feeling that CM6000 settings are stronger than CM8000 with same settings. But I >>>have no proof of course. >>> >>>Thanks, I'll "test drive" 6777 later tonight. >>> >>>Regards, >>>Jorge >>don't know what cm6777 settings are, but.. with 4 years experience, there are no >>overall settings for cm better than cm default max sel. ... however.. thewre are >>settings that are better for given games. As previously posted, the overall >>strength of the personally known chessmasters in asending order are: cm7k, cm8k, >>ckm6k, and cm55. And btw, who said that move is a "weak" move? > > >I used CM6777 for a long time, and it is definitely good. I've had less >experience with the CM8000 analog so am reserving judgement. However I feel that >increasing the selectivity to maximum (or beyond) is definitely a plus. > >The main problem I had with the CM6777 settings was that it overvalues the Queen >(in my opinion) in some situations. For example it heads for endgames of Q vs. R >+ piece + P, where the Queen can only draw. So I'm inclined to value the Queen >somewhat less, relative to that particular combination of pieces. Of course you >must also consider whether it is properly valued vs. a pair of Rooks, and vs. 3 >pieces. A compromise is required. > >I am currently using piece values which I believe are more conservative than >CM6777, but are probably an improvement over the defaults: Q=9.3, R=5, B=3.2, >N=3.1, P=1.0. As for the other settings, I use half of CM6777's deviations from >defaults: Attacker/Defender +1, Contempt for Draw +0.1, Material/Position -1, >Control of Center 101, Mobility 102, King Safety and Pawns Weakness 117. >Everything else has the default settings, except I set selectivity to 32 with a >hex editor. Call it the William Penn personality, I guess, because it's fairly >original. >WP I doubt if selectivity=32 is an advantage. I think that selectivity is about null move pruning. I am interested to know if selectivity=32 can see zunzwangs. I am interested to know if selectivity=32 can see faster some mates from my last game(I suspect that it cannot because the fastest mates is based on the idea of zungzang. For the relevant position from my last game see http://www.icdchess.com/forums/1/message.shtml?153727 It is interesting to know what is the best selectivity for solving the position. I did not try to compare selectivity=10 with other values for selectivity. Uri
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.