Search found 6948 matches

by Uri Blass
Mon Jan 11, 2021 7:02 am
Forum: Computer Chess Club: General Topics
Topic: Different scoring system for comp-comp games
Replies: 0
Views: 1885

Different scoring system for comp-comp games

The idea is the following: You play a match of 2N games from the opening position between engine A and engine B. In the first N games engine A has to win(meaning that a draw is counted as a point for B) and in the second N games engine B has to win(meaning that a draw is counted as a point for A). T...
by Uri Blass
Mon Jan 11, 2021 6:50 am
Forum: Computer Chess Club: General Topics
Topic: Do you test engines with or without using a book?
Replies: 27
Views: 5026

Re: Do you test engines with or without using a book?

Well nowadays I am working on a rating with Kings & Pawns Games. 4k3/pppppppp/8/8/8/8/PPPPPPPP/4K3 w - - 0 1 I get students playing this as part of my coaching. :P I played this when I learn chess in my childhood :) Funny but there is a lot of losses happening. Komodo is the best in Non-NNUE engine...
by Uri Blass
Sun Jan 10, 2021 10:36 pm
Forum: Computer Chess Club: General Topics
Topic: test position from lc0-stockfish
Replies: 0
Views: 1807

test position from lc0-stockfish

position is from the following game https://www.chess.com/a/3CF6AcNyC2Ra7x I was surprised to see that chess.com claims that 16.f5 of lc0 is a mistake because my stockfish claims that white has a big advantage after 16.f5 so I decided to analyze the previous moves with stockfish and it seems that 14...
by Uri Blass
Wed Jan 06, 2021 3:11 pm
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

Arasan4 It seems that Rg3 is also winning and not only Bxh7 FEN: 2rr3k/2qnbppp/p1n1p3/1p1pP3/3P1N2/1Q1BBP2/PP3P1P/1KR3R1 w - - 0 1 53/74 1:06:05 38,788,915k 9,780k +2.66 Rg1-g3 g7-g6 h2-h4 Nc6xd4 Rc1xc7 Nd4xb3 Rc7xc8 Rd8xc8 a2xb3 Be7xh4 Rg3-h3 Bh4-d8 Bd3-c2 a6-a5 Nf4-d3 Bd8-b6 Be3-h6 Nd7-b8 f3-f4 Nb...
by Uri Blass
Wed Jan 06, 2021 12:58 pm
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

Arasan1:probably correct. Arasan2:probably correct Arasan3:Maybe Nf8 also winning. Here is analysis with multi-pv=2 by stockfish 7 cores. FEN: r1bq1r1k/p1pnbpp1/1p2p3/6p1/3PB3/5N2/PPPQ1PPP/2KR3R w - - 0 1 Stockfish_20123118_x64_bmi2: 55/74 1:14:36 39,821,329k 8,895k -2.93 Be4xa8 g5-g4 Qd2-e2 g4xf3 B...
by Uri Blass
Tue Jan 05, 2021 2:47 pm
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

stockfish is extremely weak in finding Nd3 Even depth 69 and more than 5 hours on a single core did not help it to find Nd3 I did not use tablebases but I do not believe that they can make a difference here. FEN: 8/4nk2/1p3p2/1r1p2pp/1P1R1N1P/6P1/3KPP2/8 w - - 0 1 Stockfish_20123118_x64_bmi2: NNUE e...
by Uri Blass
Tue Jan 05, 2021 9:36 am
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

I looked at the history position 116 is the position I posted here I am surprised by the little progress that chess engines made in this position because already in 2011 many programs could solve it Zappa on a single core needed less than an hour to solve it even with 1 core at that time and I find ...
by Uri Blass
Tue Jan 05, 2021 8:37 am
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

default stockfish also can easily get Rxf5 FEN: 3R4/pp2r1pk/q1p3bp/2P2r2/PP6/2Q3P1/6BP/5RK1 w - - 0 1 Stockfish_20123118_x64_bmi2: NNUE evaluation using nn-62ef826d1a6d.nnue enabled 1/1 00:00 91 91k +0.85 Bg2-f3 2/3 00:00 267 134k +0.70 a4-a5 Rf5xf1+ Bg2xf1 3/4 00:00 391 196k +4.62 a4-a5 b7-b6 a5xb6...
by Uri Blass
Tue Jan 05, 2021 8:04 am
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

third position is also easy for sting single core 3R4/pp2r1pk/q1p3bp/2P2r2/PP6/2Q3P1/6BP/5RK1 w - - 0 1 Sting-sf-25: 1 00:01 258 246 +0.76 a4-a5 2 00:01 532 508 +0.76 Rf1xf5 Bg6xf5 3 00:01 799 763 +0.68 Rf1xf5 Bg6xf5 Rd8-f8 4 00:01 1k 1k +0.60 Rf1xf5 Bg6xf5 Rd8-f8 Qa6-d3 5+ 00:01 3k 3k +0.72 a4-a5 5...
by Uri Blass
Tue Jan 05, 2021 7:33 am
Forum: Computer Chess Club: General Topics
Topic: Arasan20 testsuite VS 4 top solvers
Replies: 41
Views: 6203

Re: Arasan20 testsuite VS 4 top solvers

116 is easy to solve by sting and I guess that it is also easy for GM's. I remember this position and I believe that avoiding the previous move was a harder test position. I do not understand the decision to test only stockfish clones and not to test sting-sf-25 that may be the best chess solver(not...