Author: Thorsten Czub
Date: 11:09:53 09/16/02
Go up one level in this thread
On September 16, 2002 at 12:27:49, Marc van Hal wrote: >Well what i actualy mean is Macheid looks like this >with the new methodes >[Category = Tactics] > >[Special Settings >anti-GM = 2 >combination = ON >playing style = DEFENSIVE] ?? this is maybe the macheid of christian bonnin. my macheide is different. anti GM should IMO be OFF against other computers. Combination and playing style should be normal. or ? >[Pawn Value = 94] >[Knight Value = 101] >[Bishop Value = 101] >[Rook Value = 99] >[Queen Value = 98] >[King Safety = 209] this is IMO too high for normal games against other computers. >[Mobility = 149] >[Pawn Structure = 130] >[Passed Pawns = 134] >[Pins = 150] >[Bishop Pair = 90] >[Chess Knowledge = 100] >[Attractiveness = 208] >[Attacking = 200] >[Strength of Play = 100] >[Draw Contempt Factor = 1.00] >[Selective Search = 100] >[Search Technique = DEEP] do you think deep is better than nullmove ? >[Pruning = MISC_08] * >[Pruning = MISC_15] * >[Pruning = MISC_19] * >[Pruning = MISC_20] * my style 51 uses the MISC routines 56 uses Recaptures_all thats the only difference of 56 and 51. i was sceptical too if the MISC stuff of rebel.eng is senseful. >[ANTI-GM = STRONG] should be off, or ? >[EVALUATION = NORMAL] >[Extensions (remaining)= 306] >[Extensions (checks)= 374] * >[Extensions (captures)= 76] * i do use the settings of Center / Outpost and KS-edx2 for those extensions. IMO this is better than using bit9. these 3 styles came from the alpha stage. i have example games on my web site. >Simply taken over a lot of Rebelxp for dos > >But i found out that not all prunning methodes realy did speed up the search >But not yet sure which slowed down the progress yes i know. >The ruy Lopez position is verry good for testing personelety's >Because it in fact is verry logical chess >If you increase the kingsafety it will not play Bc2 but Nbd2 because it finds an >attack on f7 after Nxb3 Qxb3. >it plays b4 because of the settings of attack >Wich is just an other example >Ofcourse programs can't see verry deep positionaly but this are simply the >reasons why it choose a certain move in that position. >Just to give an example > >Regards Marc right. this way i made the style trying different setting for months. now i am ready. cannot make it much better IMO. maybe 51 is the best, but i am not 100% sure.
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.