Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Sixteen EPD rows that need a good pounding! [Soszynski 80 test set]

Author: Ernst A. Heinz

Date: 18:03:40 03/17/00

Go up one level in this thread


Hi Dann,

>In order to create an accurate tactical suite, I would like to have these rows
>crunched by alternative software tools.  The current analysis does not get "the
>right answer" so the question is, will other programs see it?

Most of the positions you posted do *not* seem to be tactical but
rather positional in nature (e.g. "Soszynski 80 #13" is actually
from the positional subset of LCT-II).

>[D]r1r3k1/5p2/p3p1p1/b6p/Pp2N3/1P2P3/5PPP/2R1K2R w K - acd 18; acn 1259897193;
>ce -32; pv O-O Bd8 Rfd1 Be7 Kf1 Kf8 Ke1 f5 Ng3 Rc3 Rb1 Kf7 Ne2 Rc5 Rbc1 Rac8
>Rxc5 Bxc5 Nd4; pm O-O; bm Ke2; id "Soszynski 80 #8";

Positional -- "DarkThought WCCC'99" never chooses 0-0 as best. It likes
Ke2 from the beginning and even switches to Kd2 in iteration #15.

00:13:15 16.01 Kd2 Bd8 Nd6 Rc7 Rhd1 Raa7 =0.12 #297843624
>[D]r3rbk1/pp1b1p1p/2p1pBp1/2R5/3PP3/5Q2/qP2BPPP/3R2K1 w - - acd 14; acn
>1096322268; ce -45; pv Rc3 Qa5 Qf4 Qb4 Rd2 Be7 e5 Qa5 Qh4 Bf8 Rd1 Qc7 Qf4 Bg7;
>pm Rc3; bm Rh5; id "Soszynski 80 #21";

Probably positional -- "DarkThought WCCC'99" prefers Qc3 from iteration #9
onwards and scores it as nearly equal in iteration #15.

00:08:49 15.01 Qc3 a5 Qc2 b5 Rd3 b4 Qd2 a4 Rc1 a3 =-0.17 #190844581

>[D]r1b2r1k/pp2q1pp/2p2p2/2p1n2N/4P3/1PNP2QP/1PP2RP1/5RK1 w - - acd 18; acn
>422510789; ce 1; pv Nb1 Ng6 Nd2 Kg8 Nf4 Ne5 Ne2 Be6 Nc4 Bxc4 bxc4 a6 Qf4 b5 b3
>Rae8 cxb5 cxb5 Ng3; pm Nb1; bm Nd1; id "Soszynski 80 #13";

Positional -- "DarkThought WCCC'99" switches to Nd1 in iteration #13
after 3:20 min.

00:13:48 15.01 Nd1 Bd7 Ne3 Rae8 Ra1 a6 Raf1 Rf7 Nc4 =-0.19 #279492852

>[D]r5k1/4ppbp/3p1np1/qb6/4P3/1P1BBP2/P1RQ2PP/6K1 w - - acd 16; acn 205128447;
>ce -79; pv Qxa5 Rxa5 Bd2 Ne8 Bxa5 Bxd3 Rd2 Bb1 Kf2 f5 exf5 Bxf5 g4 Be6
>Rc2 Kf7 Bc3 Nf6; pm Qxa5; bm Rc8+; id "Soszynski 80 #49";

Positional -- "DarkThought WCCC'99" switches to Rc8+ in iteration #16
and scores it as roughly equal in iteration #17.

00:25:25 17.01 Rc8+ Rxc8 Qxa5 Bxd3 b4 Rc2 =-0.05 #566925263

>[D]r1b1Q3/pp3ppk/1q5p/8/8/P3P3/1PnK1PPP/3N1B1R b - - acd 11; acn 232193311; ce
>-142; pv Qd6+ Kc1 Na1 Bc4 Qg6 e4 b5 Bxf7 Nb3+ Kc2 Nd4+ Kb1 Qxg2; pm Qd6+; bm
>Bf5; id "Soszynski 80 #52";

"DarkThought WCCC'99" also likes Qd6+ from the beginning and its score
climbs steadily up to almost equal in iteration #14.

00:08:11 14.01 Qd6+ Kc1 Be6 Qb5 Rd8 Bd3+ Qxd3 Qxd3+ Rxd3 Kxc2 =-0.16 #184217202

>[D]r1b3k1/1p2b1p1/1qn1p2p/p1n1P3/3p4/P1N1BQ2/1PB2PPP/R3R1K1 w - - acd 13; acn
>179032390; ce -20; pv Qh5 dxe3 Qe8+ Bf8 Rxe3 Nd4 Bg6 Qc7 Rd1 Ncb3 Red3 Qc5 Qf7+
>Kh8 Qf4 Nc2; pm Qh5; bm Bxh6; id "Soszynski 80 #55";

"DarkThought WCCC'99" also prefers Qh5 with an ever increasing positive
score from iteration #10 onwards.

00:16:30 14.01 Qh5 dxe3 fxe3 Nd7 Rf1 Ncxe5 Qe8+ Bf8 Rxf8+ Nxf8 =1.05 #336579629

>[D]2r1r1k1/p4pp1/1p1Pn1p1/6q1/P5P1/BP2Pp1P/5P2/Q1RR2K1 b - - acd 13; acn
>178968185; ce -91; pv Rxc1 Rxc1 Rd8 Rc4 Qh4 Qe5 Qxh3 Qh2 Qxh2+ Kxh2 Nc5 Bxc5
>bxc5 Rxc5 Rxd6 Rc8+ Kh7 Rc7; pm Rxc1; bm Nf4; id "Soszynski 80 #60";

Tactical -- "DarkThought WCCC'99" switches to Nf4 in iteration #14 and
subsequently fails high on it.

00:05:07 14.01 Qh4 Qe5 Rxc1 Rxc1 Qxh3 Qg3 Qxg3+ fxg3 Rd8 Kf2 =0.42 #116364382
00:12:06 14.09 Nf4 exf4 Qh4 Rxc8 >=0.43! #268817751
01:24:48 15.01 Nf4 exf4 >=0.80! #1862062751

>[D]r2q1rk1/1b1pnpbp/p1nB2p1/1p1B4/3pP3/1N3N2/PP3PPP/R2Q1RK1 b - - acd 13; acn
>147168046; ce 7; pv h6 Nc5 Bc8 a4 bxa4 Bxc6 dxc6 e5 Bg4 Rxa4 Bxf3 Qxf3 f6 Rxd4
>fxe5; pm h6; bm Nxd5; id "Soszynski 80 #62";

Positional -- "DarkThought" wavers between various moves but also prefers
h6 in iteration #13, although it scores it as roughly -0.5.

>[D]r5k1/1p1bb1pp/1qn1prn1/p2p4/3P1PP1/3BB3/PPN3NP/R2Q1RK1 b - - acd 12; acn
>147811854; ce 19; pv Qxb2 f5 Nf8 Nf4 h6 Qe2 Nb4 Rfb1 Qc3 Nxb4 Bxb4 Rc1 Qa3 fxe6
>Bxe6 Nxe6 Nxe6; pm Qxb2; bm Bd6; id "Soszynski 80 #68";

"DarkThought WCCC'99" also prefers Qxb2 right from the beginning.

00:12:21 13.01 Qxb2 f5 Nf8 Nf4 Nb4 Nxb4 axb4 fxe6 Bxe6 Qb3 =0.66 #248590747

>[D]2r1qr1k/p5b1/2b3pp/8/5P2/1P1QB3/P4KPP/3R1B1R b - - acd 12; acn 163179495; ce
>-132; pv Be4 Qb5 Qe7 Rd2 Rc6 Bd3 a6 Bxe4 axb5 Bxc6 b4; pm Be4; bm g5; id
>"Soszynski 80 #71";

"DarkThought WCCC'99" also prefers Be4 with an ever increasing score
from iteration #9 onwards.

00:08:30 14.01  Be4 Qd7 Rc2+ Kg1 Bxg2 Qxe8 Rxe8 Bd3 Rxa2 Bxg6 =0.15 #184229079

>[D]r1b2r1k/pp1p2bp/3P4/1qpQp3/1P3p1P/4PPB1/P3N3/1R2K1R1 w - - acd 14; acn
>161319350; ce 56; pv exf4 Qc6 Qxc5 Qxc5 bxc5 exf4 Bxf4 b6 Kf2 bxc5 Rg5 c4 Be5
>Bxe5 Rxe5 Rf6 Rb4; pm exf4; bm Bxf4; id "Soszynski 80 #75";

Tactical -- "DarkThought WMCC'99" fails high on Bxf4 in iteration #12
after 1:29 min.

00:10:06 15.01 Bxf4 exf4 Rxg7 Kxg7 Qg5+ Kh8 Kf2 Qd3 Qe5+ Kg8 =2.06 #214539563

>[D]r1b1r1kb/3n1p2/p2p1npB/qp1Pp3/4P3/2N3N1/PP1QBPP1/2R1K2R w K - acd 13; acn
>143570814; ce 50; pv a3 Nf8 Qg5 N6h7 Qh4 Bf6 Qh2 b4 axb4 Qxb4 Rc2 Qd4 b3;
>pm a3; bm Nf5; id "Soszynski 80 #78";

Positional -- "DarkThought WCCC'99" shows an odd/even wavering between
a3 and Nf5 from iteration #11 onwards (up to there it sticked with Nf5).

00:00:19 10.01 Nf5 b4 Nxd6 bxc3 Qxc3 Qxc3+ Rxc3 Rd8 Nxc8 Rdxc8 =0.65 #6053696
00:00:54 11.01 Nf5 b4 Nxd6 bxc3 Rxc3 Rd8 Nxc8 Raxc8 Rxc8 Qxd2+ =0.62 #17117201
00:01:38 11.15 a3 Nb6 Nf5 Bxf5 exf5 Nbxd5 fxg6 Nxc3 Rxc3 fxg6 >=0.63! #30772345
00:05:39 12.01 a3 Nh7 Na2 Qxd2+ Bxd2 Bg7 Rc6 Bf8 Bh6 Bb7 =0.70 #105305120
00:06:03 12.02 Nf5 b4 Nxd6 bxc3 Rxc3 Rd8 Nf5 Qb4 d6 Qxc3 >=0.71! #112800509
00:08:52 13.01 Nf5 b4 Nxd6 bxc3 Rxc3 Nxe4 Nxe4 Qxa2 Nd6 Rd8 =0.73 #165288103
00:11:14 13.02 a3 Nh7 >=0.74! #208238768

>[D]r1b3k1/pp3p2/2nb2p1/6Bp/2BqpN1Q/8/PP3PPP/5RK1 w - - acd 14; acn 1138261200;
>ce -51; pv b3 Bxf4 Qxf4 Bf5 Qc7 Qd7 Qf4 Nb4 a4 Rc8 Bf6 Nd5 Bxd5 Qxd5 f3 Rc2
>fxe4 Qxe4; pm b3; bm Nxh5; id "Soszynski 80 #25";

"DarkThought WCCC'99" prefers Bb3 and scores it as slightly positive.

00:06:46 13.01 Bb3 Bxf4 Qxf4 Bf5 Qc7 Qd7 Qxd7 Bxd7 Rd1 Bf5 =0.04 #139512545

>[D]r3kb1r/pp4p1/2p1p1p1/4p3/8/2PqPQ1P/PP1N2P1/R3K2R w KQkq - acd 17; acn
>1123615068; ce 0; pv O-O-O Bc5 Nf1 Qc4 Nh2 Qxa2 Qe4 O-O Nf3 Rad8 Qxe5 Qa1+ Kc2
>Qa4+ Kc1 Bxe3+ Qxe3 Qa1+ Kc2 Qa4+ Kc1; pm O-O-O; bm Qe4; id "Soszynski 80 #27";

Positional -- "DarkThought WCCC'99" also prefers 0-0-0 right from the
beginning but scores it positively as +0.5 in iteration #16.

00:15:38 16.01 O-O-O Bc5 Qg4 Qf5 Qc4 Bxe3 Rhf1 Bf4 g3 g5 =0.50 #349875058

>[D]8/2kb4/p1p2pp1/2Pp4/3N1PP1/1P4K1/P7/8 w - - acd 17; acn 200603384; ce 0; pv
>b4 Kd8 Kf3 Ke7 g5 Be8 gxf6+ Kxf6 a4 Ke7 Kg4 Bd7+ Kg5 Be8 Kg4; pm b4; bm g5; id
>"Soszynski 80 #29";

"DarkThought WCCC'99" switches from b4 to g5 in iteration #18 and
scores it as +0.57 in iteration #19.

00:01:56 18.01 b4 Kd8 g5 Ke7 Kf2 Kf7 Kf3 Ke7 gxf6+ Kxf6 =0.39 #56431963
00:02:39 18.02 g5 fxg5 fxg5 a5 Kf4 Be8 Ke5 Bd7 >=0.40! #74532562
00:05:50 19.01 g5 f5 Kf3 a5 Ke3 Be8 Nf3 Kb7 Kd4 Kc7 =0.57 #161042153

>[D]r4r1k/bpq3pp/p1b2n2/8/2B5/PQ4NP/1P1B1PP1/R4RK1 b - - acd 11; acn 187425072;
>ce -48; pv Nd7 Rac1 Rae8 Nh1 b5 Bf7 Re2 Bd5 Bc5 Qd3 Rfe8; pm Nd7; bm Ng4; id
>"Soszynski 80 #45";

Tactical -- "DarkThought WCCC'99" fails high on Ng4 in iteration #8
after 5 sec with an ever increasing score from then onwards.

00:06:49 13.01 Ng4 hxg4 Rf3 Nf5 Rxb3 Bxb3 Bb8 g3 Qb6 Bc4 =1.62 #145388891

=Ernst=

P.S.
WWW pages of "DarkThought" at URL
http://supertech.lcs.mit.edu/~heinz/dt.



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.