Author: Mike S.
Date: 21:48:13 10/26/02
Go up one level in this thread
I had included 10 positional positions in my previous test, which I selected from 5 other testsets (2 from each). IOW, I tried to select the most convincing ones. But still, I'm not satisfied with this type of test positions. Too many doubts, i.e. it may be that the move order isn't really forced, there is a 2nd (tactical) alternative which may be even stronger, etc. (Currently I don't use positional tests anymore. Maybe that part of an engines strenght should better be judged by it's games... altough, you can't convert "judgement" into numbers like test results.) r1b2rk1/1p2qppp/1bn1p3/p3P3/8/PNQ3P1/1P1B1PBP/R4RK1 b - - bm a4; id M22#25; r5k1/p2r1bpp/2p2p2/8/n1P5/P5B1/5PPP/2R1RBK1 w - - bm c5; id M22#26; 3r1bk1/p4ppp/Qp2p3/8/1P1B4/Pq2P1P1/2r2P1P/R3R1K1 b - - bm e5; id M22#27; 7k/4K1n1/8/6P1/8/3B4/8/8 w - - bm Bg6; id M22#28; r3nrk1/ppq2pbp/2n1b1p1/4pP2/2P5/4BNN1/PP4PP/RQ2KB1R b KQ - bm e4; id M22#29; 8/p2Q2p1/1k3qbp/1pn5/2p4P/2P2P2/5BPK/5B2 w - - bm g3; id M22#30; 2r2rk1/1p1bq3/p3p2p/3pPpp1/1P1Q4/P7/2P2PPP/2R1RBK1 b - - bm Bb5; id M22#31; r1n1kb1r/1b3pp1/1qn1p3/p2pP1Bp/Qp1P1N1P/5NPB/PP3P2/2R1R1K1 w kq - bm Nxe6; id M22#32; 1r1r2k1/2b1qp1p/b1p3p1/p1p1p3/2P1P3/1PN1BP2/P1Q3PP/R2R2K1 b - - bm Rd4; id M22#33; r4k2/2p5/P1pp1n1p/2r2qp1/Q7/6BP/2P2PP1/R3R1K1 w - - bm c4; id M22#34; (M22 = Mike's Test 2.2) #28 is a tablebase position. You can decide if you want to test it without tbs.; the idea is to limit the maneuvrability of the opponent's last piece. #22 is in fact more tactical IMO. In #32, Bxe6 is probably equal to Nxe6. In #34 there'a tactical alternative with the rook maneuvre Ra1-a3-f3!, etc (the idea of the solution move c4 is to prevent Black from playing ...Rb5 and keep the b-file open for white threats directed at b7/b8. An immediate a6-a7 would be weaker.) The WM-Test(100) contains 36 positional tests. See (P) in the EPD id each (not continously numbered). http://www.computerschach.de/test/WM-Test%20(100pos).zip The download includes various data formats, and also many results for comaprison. Btw. Ruffian 1.0.0 has a very good result in the positional part of that test, among several Fritz' at the top. Regards, M.Scheidl
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.