stockfish fail high fail low

Discussion of chess software programming and technical issues.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Uri Blass
Posts: 8594
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

stockfish fail high fail low

Post by Uri Blass » Tue Apr 13, 2010 8:52 am

I wonder if stockfish 1.7 search is less accurate with a big window

I gave stockfish(1 cpu) to analyze the following position with 256 mbytes hash
The score was 89.50 at depth 35 and see what happened at depth 36

90.73
90.81(+0.08)
90.97(+0.16)
91.30(+0.33)
91.94(+0.64)
93.24(+1.30)
95.82(+2.58)
101.00(+5.18)
111.34(+10.34)
132.03(+20.69) hoping to see mate score and suddenly return to the previous
89.50

New game - Stockfish 1.7 JA, 120'/40+60'/20+30'
[D]7k/5Q2/8/8/8/3r4/2K5/8 b - - 0 1

Analysis by Stockfish 1.7 JA:

1...Rd3-g3
+- (7.07) Depth: 1 00:00:00
1...Rd3-g3 2.Qf7-f8+ Kh8-h7
+- (7.12) Depth: 2 00:00:00
1...Rd3-g3 2.Qf7-f8+ Kh8-h7 3.Qf8-f1
+- (7.12) Depth: 3 00:00:00
1...Rd3-g3 2.Qf7-f8+ Kh8-h7 3.Qf8-f2 Rg3-g8
+- (7.12) Depth: 4 00:00:00
1...Rd3-g3 2.Qf7-f8+ Kh8-h7 3.Qf8-f2 Rg3-g8 4.Kc2-d3
+- (7.22) Depth: 5 00:00:00
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Kc2-d3 Kh8-g8 4.Kd3-e4 Rg7-g2
+- (7.22) Depth: 6 00:00:00 8kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Kc2-d3 Kh8-g8 4.Kd3-e4 Rg7-f7 5.Qf6-g6+ Kg8-f8 6.Ke4-e5
+- (7.27) Depth: 7 00:00:00 13kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-h6 5.Kd3-e4 Kh6-g6 6.Ke4-e5 Kg6-h6
+- (7.27) Depth: 8 00:00:00 23kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-g6 5.Kd3-e4 Rg7-f7 6.Qf8-d8 Kg6-g7 7.Ke4-e5 Kg7-g6
+- (7.27) Depth: 9 00:00:00 44kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-g6 5.Kd3-e4 Rg7-f7 6.Qf8-d8 Kg6-g7 7.Ke4-e5 Kg7-g6 8.Qd8-h4
+- (7.27) Depth: 10 00:00:01 63kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-g6 5.Kd3-e4 Rg7-f7 6.Qf8-d8 Rf7-f1 7.Qd8-b6+ Kg6-f7 8.Qb6-b7+ Kf7-f8 9.Ke4-d5 Rf1-f7
+- (7.27) Depth: 11 00:00:01 116kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-g6 5.Kd3-e4 Rg7-f7 6.Qf8-d8 Rf7-f1 7.Qd8-b6+ Kg6-f7 8.Qb6-b7+ Kf7-f8 9.Ke4-d5 Rf1-f7 10.Qb7-b2
+- (7.27) Depth: 12 00:00:01 274kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Qf6-f8+ Kh8-h7 4.Kc2-d3 Kh7-g6 5.Kd3-e4 Rg7-f7 6.Qf8-d8 Rf7-f1 7.Qd8-b6+ Kg6-f7 8.Qb6-b7+ Kf7-f8 9.Ke4-d5 Rf1-f7 10.Qb7-c8+ Kf8-e7 11.Qc8-e6+ Ke7-f8
+- (7.27) Depth: 13 00:00:01 345kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Kc2-d3 Kh8-g8 4.Kd3-e4 Rg7-f7 5.Qf6-d4 Rf7-f8 6.Ke4-e5 Kg8-f7 7.Qd4-d7+ Kf7-g6 8.Qd7-d6+ Kg6-f7 9.Qd6-e6+ Kf7-g7 10.Qe6-c6 Rf8-g8
+- (7.32) Depth: 14 00:00:01 496kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Kc2-d3 Kh8-g8 4.Kd3-e4 Rg7-f7 5.Qf6-d4 Rf7-f8 6.Ke4-e5 Kg8-f7 7.Qd4-d7+ Kf7-g6 8.Qd7-d6+ Kg6-f7 9.Qd6-e6+ Kf7-g7 10.Qe6-e7+ Rf8-f7 11.Qe7-h4 Kg7-f8 12.Qh4-d8+ Kf8-g7 13.Ke5-e6 Kg7-g6 14.Qd8-d3+ Kg6-g7 15.Qd3-c3+
+- (7.32) Depth: 15 00:00:01 733kN
1...Rd3-g3 2.Qf7-f6+ Rg3-g7 3.Kc2-d3 Kh8-g8 4.Kd3-e4 Rg7-f7 5.Qf6-d4 Rf7-f8 6.Ke4-e5 Kg8-h7 7.Ke5-e6 Kh7-g6 8.Qd4-g4+ Kg6-h6 9.Ke6-e7 Rf8-h8 10.Ke7-f7 Rh8-h7+ 11.Kf7-f6
+- (7.37) Depth: 16 00:00:01 1180kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-d8 Ra7-f7 7.Kd4-e5 Kh7-g6 8.Qd8-e8 Kg6-g7 9.Qe8-c6 Kg7-h7 10.Ke5-e6 Rf7-f8
+- (7.32) Depth: 16 00:00:01 1567kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-d8 Ra7-a1 7.Qd8-b6 Ra1-g1 8.Kd4-e4 Rg1-f1 9.Ke4-e5 Rf1-f8 10.Qb6-c7+ Kh7-g6 11.Qc7-d6+ Kg6-g7 12.Qd6-e7+ Rf8-f7 13.Qe7-h4 Kg7-g6 14.Qh4-g4+ Kg6-h6 15.Ke5-e6 Rf7-f8 16.Ke6-e7 Rf8-h8 17.Ke7-f7
+- (7.32) Depth: 17 00:00:02 2043kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-d8 Ra7-a1 7.Qd8-b6 Ra1-g1 8.Kd4-e5 Rg1-g5+ 9.Ke5-f4 Rg5-g6 10.Qb6-c7+ Rg6-g7 11.Qc7-c2+ Kh7-g8 12.Kf4-f5 Kg8-f7 13.Qc2-b3+ Kf7-f8 14.Kf5-f6
+- (7.37) Depth: 18 00:00:02 2641kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-d8 Ra7-a1 7.Qd8-b6 Ra1-g1 8.Kd4-e5 Rg1-g5+ 9.Ke5-f4 Rg5-g6 10.Qb6-c7+ Rg6-g7 11.Qc7-c2+ Kh7-h6 12.Qc2-f5 Rg7-g8 13.Qf5-f7 Rg8-g6 14.Qf7-d5 Kh6-h7 15.Qd5-f7+ Kh7-h6
+- (7.37) Depth: 19 00:00:03 3657kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-d8 Ra7-a1 7.Qd8-b6 Ra1-g1 8.Kd4-e5 Rg1-g5+ 9.Ke5-f4 Rg5-g6 10.Qb6-c7+ Rg6-g7 11.Qc7-c2+ Kh7-h6 12.Qc2-f5 Rg7-g8 13.Qf5-f7 Rg8-g6 14.Qf7-d7
+- (7.37) Depth: 20 00:00:03 3996kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-b6 Ra7-f7 7.Kd4-e5 Rf7-f8 8.Qb6-c7+ Kh7-g6 9.Qc7-d6+ Kg6-g7 10.Qd6-e7+ Rf8-f7 11.Qe7-g5+ Kg7-h7 12.Ke5-e6 Rf7-a7 13.Qg5-h4+ Kh7-g8 14.Qh4-g3+ Ra7-g7 15.Qg3-c3 Rg7-g6+ 16.Ke6-e7 Kg8-h7 17.Ke7-f8
+- (7.42) Depth: 21 00:00:06 7873kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-b6 Ra7-f7 7.Kd4-e5 Rf7-f8 8.Qb6-c7+ Kh7-g6 9.Qc7-d6+ Kg6-g7 10.Qd6-e7+ Rf8-f7 11.Qe7-g5+ Kg7-h7 12.Ke5-e6 Rf7-a7 13.Qg5-h4+ Kh7-g8 14.Qh4-g3+ Ra7-g7 15.Qg3-c3 Rg7-g6+ 16.Ke6-e7 Rg6-g7+ 17.Ke7-f6
+- (7.42) Depth: 22 00:00:08 10597kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-b6 Ra7-f7 7.Kd4-e5 Rf7-f8 8.Qb6-c7+ Kh7-g6 9.Qc7-d6+ Kg6-g7 10.Qd6-e7+ Rf8-f7 11.Qe7-g5+ Kg7-h7 12.Ke5-e6 Rf7-a7 13.Qg5-h4+ Kh7-g8 14.Qh4-g3+ Ra7-g7 15.Qg3-c3 Rg7-g6+ 16.Ke6-e7 Rg6-g7+ 17.Ke7-f6
+- (7.42) Depth: 23 00:00:09 12976kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-a7 6.Qf6-b6 Ra7-f7 7.Kd4-e5 Rf7-f8 8.Qb6-c7+ Kh7-g6 9.Qc7-d6+ Kg6-g7 10.Qd6-e7+ Rf8-f7 11.Qe7-g5+ Kg7-h7 12.Ke5-e6 Rf7-f8 13.Qg5-e7+ Kh7-g8 14.Qe7-h4 Kg8-g7 15.Qh4-g4+ Kg7-h7 16.Ke6-e7 Rf8-g8 17.Qg4-f5+
+- (7.42) Depth: 24 00:00:11 15496kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-c2+ 3.Kb2-b3 Rc2-b2+ 4.Kb3-c3 Rb2-c2+ 5.Kc3-b4 Rc2-b2+ 6.Kb4-a3 Rb2-g2 7.Qf7-f6+ Kh8-h7 8.Qf6-h4+ Kh7-g8 9.Ka3-b4 Rg2-g6 10.Qh4-e7 Rg6-b6+ 11.Kb4-c4
+- (7.50) Depth: 25 00:00:15 22045kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Kh7-g8 6.Qf6-f5 Rg7-f7 7.Qf5-g6+ Kg8-f8 8.Qg6-h6+ Kf8-g8 9.Kd4-e5 Rf7-h7 10.Qh6-g6+ Kg8-h8 11.Qg6-g5 Rh7-g7 12.Qg5-h4+ Kh8-g8 13.Qh4-d8+ Kg8-f7 14.Ke5-f5 Rg7-h7 15.Qd8-f6+ Kf7-g8 16.Kf5-g6 Rh7-h6+ 17.Kg6xh6
+- (7.58) Depth: 25 00:00:17 25038kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (7.75) Depth: 25 00:00:21 27963kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (8.07) Depth: 25 00:00:27 35672kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (40.61) Depth: 25 00:00:27 35757kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (40.61) Depth: 25 00:00:27 35770kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (40.61) Depth: 25 00:00:28 35770kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (40.61) Depth: 25 00:00:28 35771kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (40.61) Depth: 25 00:00:29 35773kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-d7+ 6.Kd4-e5 Rd7-a7 7.Qf6-h4+ Kh7-g7 8.Qh4-d4 Ra7-e7+ 9.Ke5-d6+ Kg7-f7 10.Qd4-f4+ Kf7-e8 11.Qf4-g5
+- (49.10) Depth: 25 00:00:29 35851kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Qd4-g4+ Rh6-g6 13.Qg4-d7+ Kg7-f8 14.Ke5-f5 Rg6-h6 15.Kf5-g5 Rh6-h1 16.Qd7-d3 Kf8-g7 17.Qd3-c3+
+- (89.14) Depth: 25 00:00:52 68734kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Qd4-g4+ Rh6-g6 13.Qg4-d7+ Kg7-f8 14.Ke5-f5 Rg6-h6 15.Kf5-g5 Rh6-h1 16.Qd7-d3 Kf8-g7 17.Qd3-c3+
+- (89.14) Depth: 26 00:00:53 70582kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-f8 14.Kf5-g5 Rh6-h1 15.Qd7-c8+ Kf8-e7 16.Qc8-c7+ Ke7-e6 17.Qc7-c6+
+- (89.14) Depth: 27 00:00:56 74397kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-f8 14.Kf5-g5 Rh6-h1 15.Qd7-c8+ Kf8-e7 16.Qc8-c7+ Ke7-e6 17.Qc7-c6+
+- (89.14) Depth: 28 00:01:00 80590kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-f8 14.Kf5-g5 Rh6-h1 15.Qd7-d8+ Kf8-g7 16.Qd8-d4+ Kg7-f8 17.Qd4-b4+
+- (89.19) Depth: 29 00:01:19 106628kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-f8 14.Kf5-g5 Rh6-h1 15.Qd7-d8+ Kf8-g7 16.Qd8-d4+ Kg7-f8 17.Qd4-b4+
+- (89.19) Depth: 30 00:01:25 114599kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-f8 14.Kf5-g5 Rh6-h1 15.Qd7-d8+ Kf8-g7 16.Qd8-d4+ Kg7-f8 17.Qd4-b4+
+- (89.24) Depth: 31 00:01:42 135758kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-g8 14.Qd7-e7 Rh6-h3 15.Qe7-d6 Kg8-f7 16.Kf5-g5 Rh3-h7 17.Qd6-g6+
+- (89.24) Depth: 32 00:01:55 153511kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-f7+ Kh7-h6 7.Qf7-f8+ Kh6-h7 8.Qf8-e7+ Kh7-h6 9.Qe7-h4+ Kh6-g7 10.Kd4-e5 Rg6-b6 11.Qh4-d4 Rb6-h6 12.Ke5-f5+ Kg7-f7 13.Qd4-d7+ Kf7-g8 14.Qd7-e7 Rh6-h3 15.Qe7-d6 Rh3-h7 16.Qd6-g6+ Kg8-h8 17.Qg6-e8+
+- (89.24) Depth: 33 00:02:03 166134kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-e7+ Kh7-h6 7.Qe7-h4+ Kh6-g7 8.Kd4-e5 Rg6-c6 9.Qh4-e7+ Kg7-h6 10.Qe7-d8 Rc6-c5+ 11.Ke5-f6 Rc5-c6+ 12.Kf6-f7 Rc6-c4 13.Qd8-d2+ Kh6-h5 14.Qd2-e2+ Kh5-h4 15.Qe2xc4+ Kh4-g3 16.Kf7-f6 Kg3-g2 17.Qc4-d4
+- (90.17) Depth: 34 00:02:24 195157kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-c3 Kh8-h7 5.Kc3-d4 Rg7-g6 6.Qf6-e7+ Rg6-g7 7.Qe7-h4+ Kh7-g6 8.Kd4-e5 Rg7-f7 9.Qh4-g4+ Kg6-h7 10.Qg4-e4+ Kh7-h8 11.Ke5-d6 Rf7-f6+ 12.Kd6-e7 Rf6-f7+ 13.Ke7-e6 Rf7-f8 14.Qe4-e5+ Kh8-g8 15.Ke6-e7 Rf8-f7+ 16.Ke7-d6 Kg8-h7 17.Kd6-e6
+- (90.25) Depth: 34 00:02:43 218340kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-c2+ 3.Kb2-b3 Rc2-b2+ 4.Kb3-a4 Rb2-b4+ 5.Ka4-a3 Rb4-a4+ 6.Ka3-b2 Ra4-a5 7.Qf7-d7 Ra5-g5 8.Kb2-c3 Rg5-g7 9.Qd7-e8+ Kh8-h7 10.Qe8-h5+ Kh7-g8 11.Kc3-c4 Rg7-a7 12.Qh5-e5 Ra7-b7 13.Qe5-d5+ Rb7-f7 14.Kc4-c5 Kg8-g7 15.Qd5-g5+ Kg7-h7 16.Kc5-d6 Rf7-f1 17.Qg5-h4+
+- (90.41) Depth: 34 00:02:51 229114kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-c2+ 3.Kb2-b3 Rc2-b2+ 4.Kb3-a4 Rb2-b4+ 5.Ka4-a3 Rb4-a4+ 6.Ka3-b2 Ra4-a5 7.Qf7-d7 Ra5-g5 8.Kb2-c3 Rg5-g7 9.Qd7-e8+ Kh8-h7 10.Qe8-h5+ Kh7-g8 11.Kc3-c4 Rg7-a7 12.Qh5-e5 Ra7-b7 13.Qe5-d5+ Rb7-f7 14.Kc4-c5 Kg8-g7 15.Qd5-g5+ Kg7-h7 16.Kc5-d6 Rf7-f1 17.Qg5-h4+
+- (89.50) Depth: 34 00:03:06 251423kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-c2+ 3.Kb2-b3 Rc2-b2+ 4.Kb3-a4 Rb2-b4+ 5.Ka4-a3 Rb4-a4+ 6.Ka3-b2 Ra4-a5 7.Qf7-d7 Ra5-g5 8.Kb2-c3 Rg5-g7 9.Qd7-e8+ Kh8-h7 10.Qe8-h5+ Kh7-g8 11.Kc3-c4 Rg7-a7 12.Qh5-e5 Ra7-b7 13.Qe5-d5+ Rb7-f7 14.Kc4-c5 Kg8-g7 15.Qd5-g5+ Kg7-h7 16.Kc5-d6 Rf7-f1 17.Qg5-h4+
+- (89.50) Depth: 35 00:03:28 282079kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-c2+ 3.Kb2-b3 Rc2-b2+ 4.Kb3-a4 Rb2-b4+ 5.Ka4-a3 Rb4-a4+ 6.Ka3-b2 Ra4-a5 7.Qf7-d7 Ra5-g5 8.Kb2-c3 Rg5-g7 9.Qd7-e8+ Kh8-h7 10.Qe8-h5+ Kh7-g8 11.Kc3-c4 Rg7-a7 12.Qh5-e5 Ra7-b7 13.Qe5-d5+ Rb7-f7 14.Kc4-c5 Kg8-g7 15.Qd5-g5+ Kg7-f8 16.Kc5-d6 Rf7-g7 17.Qg5-f6+
+- (90.73) Depth: 36 00:03:54 314263kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-h4+ Rg6-h6 7.Qh4-e4+ Kh7-g7 8.Qe4-e7+ Kg7-g6 9.Kc4-d5 Rh6-h5+ 10.Kd5-e4 Rh5-h1 11.Qe7-e6+ Kg6-h7 12.Qe6-f5+ Kh7-g7 13.Qf5-g4+ Kg7-h6 14.Qg4-f4+ Kh6-g6 15.Qf4-g3+ Kg6-h7 16.Ke4-f5 Rh1-h6 17.Qg3-g5
+- (90.81) Depth: 36 00:04:10 335680kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g8 6.Qf6-f7+ Kh7-h8 7.Qf7-e7 Rg8-g7 8.Qe7-e5 Kh8-h7 9.Qe5-h5+ Kh7-g8 10.Kc4-d5 Rg7-g3 11.Qh5-e8+ Kg8-h7 12.Qe8-e7+ Rg3-g7 13.Qe7-h4+ Kh7-g6 14.Kd5-e6 Rg7-g8 15.Qh4-f6+ Kg6-h7 16.Qf6-e7+ Rg8-g7 17.Qe7-h4+
+- (90.97) Depth: 36 00:04:23 353257kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g8 6.Qf6-f7+ Kh7-h8 7.Qf7-e7 Rg8-g2 8.Qe7-e5+ Kh8-h7 9.Qe5-e4+ Rg2-g6 10.Kc4-d5 Kh7-g7 11.Qe4-e7+ Kg7-h8 12.Qe7-h4+ Kh8-g7 13.Kd5-e5 Rg6-g2 14.Qh4-f6+ Kg7-h7 15.Qf6-f5+ Rg2-g6 16.Qf5-d7+ Rg6-g7 17.Qd7-d3+
+- (91.30) Depth: 36 00:04:33 367510kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g8 6.Qf6-f7+ Kh7-h8 7.Qf7-e7 Rg8-g6 8.Kc4-d5 Rg6-g8 9.Kd5-e5 Rg8-g2 10.Qe7-d8+ Rg2-g8 11.Qd8-d4 Rg8-f8 12.Ke5-e6+ Kh8-g8 13.Qd4-h4 Kg8-g7 14.Qh4-e7+ Kg7-g8 15.Qe7-d7 Rf8-f3 16.Qd7-d8+ Rf3-f8 17.Qd8-e7
+- (91.94) Depth: 36 00:04:35 370258kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Kh7-g8 6.Kc4-d5 Rg7-f7 7.Qf6-g5+ Rf7-g7 8.Qg5-e5 Kg8-f8 9.Qe5-f6+ Rg7-f7 10.Qf6-d8+ Kf8-g7 11.Qd8-g5+ Kg7-h7 12.Kd5-d6 Rf7-f1 13.Qg5-h4+ Kh7-g6 14.Kd6-e6 Rf1-f8 15.Qh4-g4+ Kg6-h7 16.Qg4-h5+ Kh7-g7 17.Qh5-g5+
+- (93.24) Depth: 36 00:04:36 371419kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-c7+ 6.Kc4-d4 Rc7-a7 7.Qf6-d6 Ra7-g7 8.Kd4-e5 Rg7-g5+ 9.Ke5-f4 Rg5-g7 10.Qd6-d3+ Kh7-g8 11.Qd3-d8+ Kg8-h7 12.Kf4-f5 Rg7-f7+ 13.Kf5-e6 Rf7-a7 14.Qd8-d3+ Kh7-g8 15.Qd3-g6+ Kg8-h8 16.Qg6-h5+ Kh8-g7 17.Qh5-e5+
+- (95.82) Depth: 36 00:04:38 373675kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-f4 Rg6-a6 7.Kc4-d5 Ra6-a5+ 8.Kd5-e6 Ra5-a6+ 9.Ke6-f5 Ra6-a5+ 10.Kf5-g4 Ra5-a6 11.Qf4-e4+ Kh7-h6 12.Qe4-e3+ Kh6-g7 13.Qe3-c3+ Kg7-f7 14.Qc3-c4+ Kf7-e7 15.Qc4xa6 Ke7-d7 16.Qa6-a7+ Kd7-e8 17.Qa7-b6
+- (101.00) Depth: 36 00:04:39 375212kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-f4 Rg6-a6 7.Kc4-d5 Ra6-a5+ 8.Kd5-e6 Ra5-a6+ 9.Ke6-f5 Ra6-a5+ 10.Kf5-g4 Ra5-a6 11.Qf4-e4+ Kh7-h6 12.Qe4-e3+ Kh6-g7 13.Qe3-c3+ Kg7-f7 14.Qc3-c4+ Kf7-e7 15.Qc4xa6 Ke7-d7 16.Qa6-a7+ Kd7-e8 17.Qa7-b6
+- (111.34) Depth: 36 00:04:47 382665kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-f4 Rg6-a6 7.Kc4-d5 Ra6-a5+ 8.Kd5-e6 Ra5-a6+ 9.Ke6-f5 Ra6-a5+ 10.Kf5-g4 Ra5-a6 11.Qf4-e4+ Kh7-h6 12.Qe4-e3+ Kh6-g7 13.Qe3-c3+ Kg7-f7 14.Qc3-c4+ Kf7-e7 15.Qc4xa6 Ke7-d7 16.Qa6-a7+ Kd7-e8 17.Qa7-b6
+- (132.03) Depth: 36 00:04:47 383587kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-f4 Rg6-a6 7.Kc4-d5 Ra6-a5+ 8.Kd5-e6 Ra5-a6+ 9.Ke6-f5 Ra6-a5+ 10.Kf5-g4 Ra5-a7 11.Qf4-h2+ Kh7-g6 12.Qh2-h5+ Kg6-g7 13.Qh5-e5+ Kg7-h7 14.Kg4-g5 Ra7-f7 15.Qe5-d4 Rf7-c7 16.Qd4-e4+ Kh7-g8 17.Qe4-e8+
+- (89.50) Depth: 36 00:06:48 543376kN
1...Rd3-c3+ 2.Kc2-b2 Rc3-g3 3.Qf7-f6+ Rg3-g7 4.Kb2-b3 Kh8-h7 5.Kb3-c4 Rg7-g6 6.Qf6-f4 Rg6-a6 7.Kc4-d5 Ra6-a5+ 8.Kd5-e6 Ra5-a6+ 9.Ke6-f5 Ra6-a5+ 10.Kf5-g4 Ra5-a7 11.Qf4-h2+ Kh7-g6 12.Qh2-h5+ Kg6-g7 13.Qh5-e5+ Kg7-h7 14.Kg4-g5 Ra7-f7 15.Qe5-d4 Rf7-c7 16.Qd4-e4+ Kh7-g8 17.Qe4-e8+
+- (89.50) Depth: 37 00:14:43 1140549kN

(, 13.04.2010)

zamar
Posts: 613
Joined: Sun Jan 18, 2009 6:03 am

Re: stockfish fail high fail low

Post by zamar » Tue Apr 13, 2010 9:30 am

Zugzwangs play major role in KQKR ending. Because zugzwang detection is disabled by default (this is a bug, but we couldn't fix it without changing functionality), you can see strange results like this.
Joona Kiiski

zamar
Posts: 613
Joined: Sun Jan 18, 2009 6:03 am

Re: stockfish fail high fail low

Post by zamar » Tue Apr 13, 2010 4:31 pm

Sorry, this is not a zugzwang issue at all, but related to fact that TT scores are not allowed at PV nodes.
Joona Kiiski

Uri Blass
Posts: 8594
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: stockfish fail high fail low

Post by Uri Blass » Thu Apr 15, 2010 3:37 am

I wonder what is the reason for not allowing them?

Are you afraid from draws by the 50 move rule or draws by repetition or that you can score as wins or maybe there are more reasons?

If the reason is draws you can disable tranposition tables only when the score before the fail high suggest good chances to win(otherwise draws are not a big risk).

It is clearly annoying to see big fail high in analyzing correspondecne games when there is a fail low after them when I do not know if to trust the fail high and the main question is if there are practical cases when the fail high is really wrong(not because of repetition or 50 move rule because a repetition is irrelevant when the position never repeated twice when 50 move rule is also probably irrelevant because stockfish needs to calculate a line of more than 90 plies with no capture and no pawn move for it).

Uri

zamar
Posts: 613
Joined: Sun Jan 18, 2009 6:03 am

Re: stockfish fail high fail low

Post by zamar » Thu Apr 15, 2010 4:22 am

Uri Blass wrote:I wonder what is the reason for not allowing them?

Are you afraid from draws by the 50 move rule or draws by repetition or that you can score as wins or maybe there are more reasons?
Also three-fold repetition and the fact that in non-PV nodes we prune much more aggressively than in PV-nodes, so they do not contain accurate enough info for PV nodes.
If the reason is draws you can disable tranposition tables only when the score before the fail high suggest good chances to win(otherwise draws are not a big risk).
You've suggested a lot of different kind of "hacks" to fix fail high/low-problems. The problem with those is that ruin the codebase and make further improvements harder to implement. So, there need to be a demonstration of a big boost in elo strength, before we can even consider sth like that :(
It is clearly annoying to see big fail high in analyzing correspondecne games when there is a fail low after them when I do not know if to trust the fail high and the main question is if there are practical cases when the fail high is really wrong(not because of repetition or 50 move rule because a repetition is irrelevant when the position never repeated twice when 50 move rule is also probably irrelevant because stockfish needs to calculate a line of more than 90 plies with no capture and no pawn move for it).
I understand, but we need a clean solution to fix this. About the practical cases: I don't know :(
Joona Kiiski

mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 7:17 pm

Re: stockfish fail high fail low

Post by mcostalba » Thu Apr 15, 2010 7:59 am

zamar wrote:I understand, but we need a clean solution to fix this. About the practical cases: I don't know :(
Joona, I think we should give a kind of "clean solution" definition to our friends here ;-)

Clean solution: a solution is considered clean _only_ when

1) It removes more lines of code then it adds

2) Is supported by strong and serious testing evidence


It is interesting to note that Uri's suggestions although always imaginative and creative consistently fail both the above points. I have since long abandoned the wishful thinking of point 2, but at least point 1 _could_ be honored once in a while.

Uri, just to be clear. Read my lips: "No Hacks !" :-)

The secret of SF quick improvments is that we try our best to stay away from tricks, hacks, shortcuts and all that kind of garbage because they are no way out roads, you go along that line hoping it brings you sometime and suddendly you find out the engine is a mess and you cannot no more progress on that.

If implementation is simple and clear it can be improved further, otherwise you are at the end of the road.

User avatar
michiguel
Posts: 6388
Joined: Thu Mar 09, 2006 7:30 pm
Location: Chicago, Illinois, USA
Contact:

Re: stockfish fail high fail low

Post by michiguel » Thu Apr 15, 2010 3:06 pm

mcostalba wrote:
zamar wrote:I understand, but we need a clean solution to fix this. About the practical cases: I don't know :(
Joona, I think we should give a kind of "clean solution" definition to our friends here ;-)

Clean solution: a solution is considered clean _only_ when

1) It removes more lines of code then it adds

2) Is supported by strong and serious testing evidence


It is interesting to note that Uri's suggestions although always imaginative and creative consistently fail both the above points. I have since long abandoned the wishful thinking of point 2, but at least point 1 _could_ be honored once in a while.

Uri, just to be clear. Read my lips: "No Hacks !" :-)

The secret of SF quick improvments is that we try our best to stay away from tricks, hacks, shortcuts and all that kind of garbage because they are no way out roads, you go along that line hoping it brings you sometime and suddendly you find out the engine is a mess and you cannot no more progress on that.

If implementation is simple and clear it can be improved further, otherwise you are at the end of the road.
This is off-topic, but I wanted to mention this for while.

Over the years, learned that a computer scientist define clean code as everything that looks like his/her own code :-)

Clean, elegant, beautiful, ugly are words imported from other disciplines that do not necessarily apply to comp. sci. They are completely subjective. What is elegant for somebody, it could be an unnecessary trick, or even ineffective for somebody else.

Words that are more useful (and a little bit less subjective, but not perfect) to describe code are "straightforward", "readable", "maintainable", "portable". I can attempt to define those, but any definition of "clean" is doom to failure. You cannot attach virtue or vice to the number of lines of code, IMHO.

Miguel

User avatar
Eelco de Groot
Posts: 4162
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: stockfish fail high fail low

Post by Eelco de Groot » Thu Apr 15, 2010 3:54 pm

Code: Select all

// Impromptu 

#if !defined elegance

Like a poem or song.

You can see if it is wrong.

When the words are in their place.

And all lines are in one piece.

When the comments are short.

When you can do a Linux port!

When you keep reading through the night.

Can't put it down, turn off the light.

You know that it is right!

#endif

[D]7k/5Q2/8/8/8/3r4/2K5/8 b - -

Engine: Rainbow Serpent 1.6.3s(dc) Build 286 (Athlon 2009 MHz, 256 MB)
by Tord Romstad, Marco Costalba, Joona Kiiski Modifications: Dann Corbit

1.00 0:00 -7.19 1...Rg3 (141) 0

2.00 0:00 -7.24 1...Rg3 2.Qf8+ Kh7 (270) 0

3.00 0:00 -7.29 1...Rg3 2.Qf6+ Rg7 3.Kd3 (13.234) 40

4.00 0:00 -7.34 1...Rg3 2.Qf6+ Rg7 3.Qf8+ Kh7 4.Kd3 Rg8
5.Qf5+ Kg7 6.Kd4 (19.522) 59

5.00 0:00 -7.34 1...Rg3 2.Qf6+ Rg7 3.Kd3 Kg8 4.Kd4 Rh7
5.Qg6+ Rg7 6.Qe6+ Kf8 7.Kd5 (50.263) 140

6.00 0:00 -7.34 1...Rg3 2.Qf6+ Rg7 3.Kd3 Kg8 4.Kd4 Rh7
5.Qg6+ Rg7 6.Qe6+ Kf8 7.Kd5 (94.532) 232

7.00 0:00 -7.42 1...Rg3 2.Qf6+ Rg7 3.Qf8+ Kh7 4.Kd3 Rg5
5.Ke4 Kg6 6.Qd8 Kh6 7.Qf6+ Rg6 (172.656) 368

7.00 0:00 -7.29 1...Rd2+ 2.Kc3 Rc2+ 3.Kd3 Rd2+ 4.Ke3 Re2+
5.Kd4 Rd2+ 6.Ke5 Re2+ 7.Kd5 Rd2+
8.Ke6 Re2+ 9.Kd7 Rd2+ 10.Kc6 Rc2+
11.Kb7 Rh2 (567.265) 698

8.00 0:00 -7.29 1...Rd2+ 2.Kc3 Rc2+ 3.Kd3 Rd2+ 4.Kc4 Rc2+
5.Kb5 Rb2+ 6.Ka4 Rg2 7.Qh5+ Kg7
8.Kb3 Rg6 9.Kc4 Ra6 10.Qg5+ Kh7
11.Kd4 Rg6 12.Qh5+ Rh6 13.Qf5+ Kg8
14.Ke5 (633.265) 723

9.00 0:00 -7.29 1...Rd2+ 2.Kc3 Rc2+ 3.Kd3 Rd2+ 4.Kc4 Rc2+
5.Kb5 Rb2+ 6.Ka4 Rg2 7.Qh5+ Kg7
8.Kb3 Rg6 9.Kc4 Ra6 10.Qg5+ Kh7
11.Kd4 Rg6 12.Qh5+ Rh6 13.Qf5+ Kg8
14.Ke5 (662.273) 730

10.00 0:01 -7.35 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kd5 Rd2+ 6.Kc4 Rc2+ 7.Kb3 Rb2+
8.Ka3 Rg2 9.Qh5+ Kg7 10.Qe5+ Kh7
11.Qe4+ Rg6 12.Kb3 Kh6 13.Kc4 Ra6
14.Kd5 Kg7 (1.843.238) 1008

11.01 0:02 -7.37 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb6 Rb2+ 6.Ka7 Rg2 7.Qh5+ Kg7
8.Kb6 Rg6+ 9.Kc5 Kf6 10.Kd6 Rg5
11.Qh3 Rg6 12.Qf3+ Kg5+ 13.Ke5 Rh6 (2.690.039) 1062

12.01 0:03 -7.42 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb6 Rb2+ 6.Ka7 Rg2 7.Qh5+ Kg7
8.Kb6 Rg6+ 9.Kc5 Rh6 10.Qg5+ Kh7
11.Qf5+ Rg6 12.Kd5 Kh6 13.Ke4 Ra6
14.Qf8+ Kg5 (3.639.629) 1114

13.01 0:07 -7.42 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb4 Rb2+ 6.Kc4 Rc2+ 7.Kb3 Rb2+
8.Ka4 Rb4+ 9.Ka3 Rb6 10.Qe7 Rg6
11.Kb4 Rg7 12.Qe8+ Kh7 13.Qh5+ Kg8
14.Kc5 Rh7 (9.250.962) 1213

14.01 0:08 -7.42 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb4 Rb2+ 6.Kc4 Rc2+ 7.Kb3 Rb2+
8.Ka4 Rb4+ 9.Ka3 Rb6 10.Qe7 Rg6
11.Kb3 Rg7 12.Qf8+ Kh7 13.Kc3 Rg8
14.Qf7+ Rg7 (10.457.283) 1237

15.01 0:09 -7.43 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb4 Rb2+ 6.Ka5 Rb5+ 7.Ka4 Rb4+
8.Ka3 Rb6 9.Qe7 Rg6 10.Kb4 Rg7
11.Qf8+ Kh7 12.Kc4 Rg5 13.Kd4 Kg6
14.Ke4 Kh5 (12.087.731) 1251

16.01 0:16 -7.47 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb6 Rb2+ 6.Ka7 Rg2 7.Qh5+ Kg7
8.Qe5+ Kh7 9.Kb6 Rg6+ 10.Kc5 Kh6
11.Qe3+ Kh5 12.Kd5 Rg5+ 13.Ke6 Kg4
14.Qe4+ Kh3 (21.601.121) 1312

17.01 0:21 -7.47 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb6 Rb2+ 6.Ka7 Rg2 7.Qh5+ Kg7
8.Qe5+ Kh7 9.Kb6 Rg6+ 10.Kc5 Kh6
11.Qe3+ Kh5 12.Kd5 Rg5+ 13.Ke6 Kg4
14.Qe4+ Kh3 (28.499.266) 1328

18.01 0:26 -7.47 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb6 Rb2+ 6.Ka7 Rg2 7.Qh5+ Kg7
8.Qe5+ Kh7 9.Kb6 Rg6+ 10.Kc5 Kh6
11.Qe3+ Kh5 12.Kd5 Kg4 13.Qe4+ Kh5
14.Ke5 Kg5 (34.999.000) 1345

19.01 0:37 -7.67-- 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Ka4 Rg2 7.Qh5+ Kg7
8.Qe5+ Kh7 9.Kb4 Rg6 10.Qh5+ Kg7
11.Kc5 Kf6 12.Kd6 Kf7+ 13.Ke5 Kg7
14.Qf5 Kh6 (50.365.843) 1328

20.01 0:45 -7.67 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Ka4 Rg2 7.Qh5+ Kg7
8.Qe5+ Kh7 9.Kb4 Rg6 10.Qh5+ Kg7
11.Kc5 Kf6 12.Kd6 Kf7+ 13.Ke5 Kg7
14.Qf5 Kh6 (62.703.546) 1380

21.01 1:02 -8.48-- 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Kc6 Rg2 7.Qe8+ Kg7
8.Qe5+ Kh7 9.Qe4+ Rg6+ 10.Kd5 Kh6
11.Qh4+ Kg7 12.Ke5 Rh6 13.Qg5+ Kh7
14.Kf5 Ra6 (86.764.627) 1393

22.01 1:28 -11.91-- 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rb2 4.Qf5 Kg8
5.Kc5 (121.730.503) 1375

23.01 3:06 -25.65-- 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc4 Rc2+
5.Kb3 Rb2+ 6.Ka3 Rb6 7.Qe7 Rg6 8.Kb3 Rg7
9.Qe5 Kg8 10.Kc4 Rg6 11.Kd5 Rh6
12.Qg5+ Kh7 13.Ke5 Rh2 14.Qg3 Rb2 (261.715.681) 1404

24.01 5:38 -25.66 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Kc6 Rc2+ 7.Kb7 Rb2+
8.Ka8 Rb6 9.Qf8+ Kh7 10.Qf2 Rg6
11.Qh4+ Rh6 12.Qe4+ Rg6 13.Kb8 Kh6
14.Qh4+ Kg7 (492.052.150) 1455

25.01 10:52 -25.66 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Kc6 Rc2+ 7.Kb7 Rb2+
8.Ka8 Rb6 9.Qf8+ Kh7 10.Qf2 Rg6
11.Qh4+ Rh6 12.Qe4+ Kg7 13.Kb7 Kf6
14.Qf4+ Kg6 (968.679.792) 1484

26.01 13:10 -39.39-- 1...Rd2+ 2.Kc3 Rc2+ 3.Kd4 Rd2+ 4.Kc5 Rc2+
5.Kb5 Rb2+ 6.Kc6 Rb6+ 7.Kc5 Rb5+
8.Kc4 Rb4+ 9.Kc3 Rb6 10.Qh5+ Kg7
11.Qe5+ Rf6 12.Kd4 Kg6 13.Qg3+ Kf7
14.Ke5 Re6+ (1.167.295.683) 1476

27.01 22:05 -94.32-- 1...Rd2+ 2.Kc3 Rh2 3.Qf8+ Kh7 4.Qe7+ Kg8
5.Qd8+ Kf7 6.Qc7+ Ke6 7.Qxh2 Kd5
8.Qf4 Kc6 9.Qe5 Kb7 10.Kc4 Kc8
11.Qe8+ Kc7 12.Kd5 Kb6 13.Qe3+ Kb5
14.Qb3+ Ka5 (1.967.470.725) 1484

28.01 279:45 -M52 1...Rd2+ 2.Kc3 Rc2+ 3.Kb4 Rb2+ 4.Ka4 Rg2
5.Qh5+ Kg7 6.Qe5+ Kh7 7.Kb3 Rg8
8.Qh5+ Kg7 9.Kc4 Kf6 10.Qf3+ Ke7
11.Qb7+ Kf8 12.Qa8+ Kg7 13.Qd5 Rf8
14.Qg5+ Kh7 (23.365.584.486) 1392


best move: Rd3-d2 time: 313:24.047 min n/s: 1.371.055 nodes: 25.781.370.305

Eelco
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 7:17 pm

Re: stockfish fail high fail low

Post by mcostalba » Thu Apr 15, 2010 5:49 pm

michiguel wrote: You cannot attach virtue or vice to the number of lines of code, IMHO.
Hi Miguel,

improving strength you know is not easy, improving strength reducing lines of code is even more difficult and _normally_ it means you have a clear understanding of the problem and of all its side-effects: when this conditions is true your code _happens_ to be also clean and elegant without any visible effort.

It is a consequence of having a clean and complete vision of the problem.

When your code is a mess it means you have not understood the big picture and you are blindly handwaving here and there.

This is IMHO, of course ;-)

jwes
Posts: 778
Joined: Sat Jul 01, 2006 5:11 am

Re: stockfish fail high fail low

Post by jwes » Thu Apr 15, 2010 9:23 pm

michiguel wrote: This is off-topic, but I wanted to mention this for while.

Over the years, learned that a computer scientist define clean code as everything that looks like his/her own code :-)

Clean, elegant, beautiful, ugly are words imported from other disciplines that do not necessarily apply to comp. sci. They are completely subjective. What is elegant for somebody, it could be an unnecessary trick, or even ineffective for somebody else.

Words that are more useful (and a little bit less subjective, but not perfect) to describe code are "straightforward", "readable", "maintainable", "portable". I can attempt to define those, but any definition of "clean" is doom to failure. You cannot attach virtue or vice to the number of lines of code, IMHO.

Miguel
We may well disagree on whether some code that neither of us has written is clear or elegant, but if I find some code that I have written ugly or awkward, I think it likely you would agree with me. I believe that elegant has an objective meaning, but it is not easily defined. It is like trying to define "quality" in the context of source code. Since truly elegant code is so rare the word is frequently misused.

Post Reply