[D] 8/6p1/8/p1p1p1kP/1p1pP1P1/1P1P3K/1PP5/8 b - -
This is from Havasi-Peko, Budapest 1976 - it is one of the examples in Dvoretsky's endgame book.
bm is c4, which I have verified at long time control. Arasan gets this but not rapidly.
--Jon
difficult endgame test
Moderators: hgm, Rebel, chrisw
-
jdart
- Posts: 4366
- Joined: Fri Mar 10, 2006 5:23 am
- Location: http://www.arasanchess.org
-
Steve Maughan
- Posts: 1221
- Joined: Wed Mar 08, 2006 8:28 pm
- Location: Florida, USA
Re: difficult endgame test
Hmmmmmm, Rybka Beta 1 gets it... and so does Strelka. Both have amazingly similar scores and PVs with a two ply gap. I really don't think Strelka is a clone but there's something fishy (bad pun - sorry!) or at the very least there's some link between Strelka and Rybka 1 that's not public.
Benefit of the doubt regards,
Steve
P.S. The next ply also gave a score of +0.78 for both engines
[D]8/6p1/8/p1p1p1kP/1p1pP1P1/1P1P3K/1PP5/8 b - -
Engine: Rybka 1.0 Beta 32-bit (512 MB)
by Vasik Rajlich
20.01 0:11 0.00 1...Kf4 2.Kh4 c4 3.g5 a4 4.h6 axb3
5.cxb3 cxd3 6.hxg7 (5.082.368) 462
20.02 0:26 +0.18 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (7.152.598) 275
21.01 0:33 +0.19 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (8.716.069) 272
22.01 0:49 +0.24 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (11.937.317) 248
23.01 1:19 +0.34 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (17.679.155) 229
24.01 3:02 +0.72 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (30.951.922) 173
25.01 5:15 +0.78 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (48.798.671) 158
[D]8/6p1/8/p1p1p1kP/1p1pP1P1/1P1P3K/1PP5/8 b - -
Engine: Strelka 1.8 UCI (512 MB)
by Jury Osipov
22.06 0:12 0.00 1...Kf4 2.Kh4 c4 3.g5 a4 4.h6 axb3
5.cxb3 cxd3 6.hxg7 d2 7.g8Q d1Q
8.Qf7+ Kxe4 9.g6 Qe1+ 10.Kg5 d3
11.g7 Qg3+ 12.Kf6 Qf2+ 13.Kg6 Qg3+
14.Kf6 (22.239.310) 1806
23.06 0:44 +0.13 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (78.468.239) 1766
24.06 0:56 +0.24 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (100.222.335) 1766
25.06 1:24 +0.34 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (148.575.610) 1764
26.06 2:45 +0.72 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (289.642.612) 1747
27.06 4:57 +0.78 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (517.720.533) 1742
Benefit of the doubt regards,
Steve
P.S. The next ply also gave a score of +0.78 for both engines
[D]8/6p1/8/p1p1p1kP/1p1pP1P1/1P1P3K/1PP5/8 b - -
Engine: Rybka 1.0 Beta 32-bit (512 MB)
by Vasik Rajlich
20.01 0:11 0.00 1...Kf4 2.Kh4 c4 3.g5 a4 4.h6 axb3
5.cxb3 cxd3 6.hxg7 (5.082.368) 462
20.02 0:26 +0.18 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (7.152.598) 275
21.01 0:33 +0.19 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (8.716.069) 272
22.01 0:49 +0.24 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (11.937.317) 248
23.01 1:19 +0.34 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (17.679.155) 229
24.01 3:02 +0.72 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (30.951.922) 173
25.01 5:15 +0.78 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 (48.798.671) 158
[D]8/6p1/8/p1p1p1kP/1p1pP1P1/1P1P3K/1PP5/8 b - -
Engine: Strelka 1.8 UCI (512 MB)
by Jury Osipov
22.06 0:12 0.00 1...Kf4 2.Kh4 c4 3.g5 a4 4.h6 axb3
5.cxb3 cxd3 6.hxg7 d2 7.g8Q d1Q
8.Qf7+ Kxe4 9.g6 Qe1+ 10.Kg5 d3
11.g7 Qg3+ 12.Kf6 Qf2+ 13.Kg6 Qg3+
14.Kf6 (22.239.310) 1806
23.06 0:44 +0.13 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (78.468.239) 1766
24.06 0:56 +0.24 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (100.222.335) 1766
25.06 1:24 +0.34 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (148.575.610) 1764
26.06 2:45 +0.72 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (289.642.612) 1747
27.06 4:57 +0.78 1...c4 2.bxc4 a4 3.c5 a3 4.bxa3 bxa3
5.c6 a2 6.c7 a1Q 7.c8Q Qf1+ 8.Kg3 Qf4+
9.Kh3 Qf3+ 10.Kh2 Qf2+ 11.Kh3 Qh4+
12.Kg2 Qxg4+ 13.Qxg4+ Kxg4 14.Kf2 Kxh5 (517.720.533) 1742
-
bob
- Posts: 20943
- Joined: Mon Feb 27, 2006 7:30 pm
- Location: Birmingham, AL
Re: difficult endgame test
I ran this with crafty... it finds this move at depth=11, .1 seconds, on my core2 duo laptop. by depth=18, .5 seconds, the score is in black's favor by 1/2 pawn.... by depth 21, it is 1.5 pawns in blacks favor, taking about 2 seconds...
Code: Select all
11 0.19 2.59 1. ... g6 2. hxg6 Kxg6 3. Kh4 Kf7 4.
g5 Kg7 5. Kh5 Kf7 6. g6+ Kg7 7. Kg5
(s=3)
11 0.23 -1 1. ... c4!!
11 0.24 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 Kf4 6. c7 a2 7. c8=Q a1=Q
11-> 0.24 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 Kf4 6. c7 a2 7. c8=Q a1=Q
(s=3)
12 0.24 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 Kf4 6. c7 a2 7. c8=Q a1=Q
(s=2)
12-> 0.27 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 Kf4 6. c7 a2 7. c8=Q a1=Q
(s=2)
13 0.31 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Kf4
13-> 0.31 1.68 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Kf4
(s=2)
14 0.34 -1 1. ... c4!!
14 0.35 1.00 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qh1+
8. Kg3 Qh4+ 9. Kf3 Qh3+ 10. Kf2 Qxg4
11. Qxg4+ Kxg4
14-> 0.39 1.00 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qh1+
8. Kg3 Qh4+ 9. Kf3 Qh3+ 10. Kf2 Qxg4
11. Qxg4+ Kxg4 (s=2)
15 0.42 -1 1. ... c4!!
15 0.45 0.16 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. h6 gxh6 6. c6 a2 7. c7 a1=Q
8. c8=Q Kf4
15-> 0.51 0.16 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. h6 gxh6 6. c6 a2 7. c7 a1=Q
8. c8=Q Kf4
16 0.57 -0.01 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qh1+
8. Kg3 Qg1+ 9. Kh3 Qh1+
16-> 0.60 -0.01 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qh1+
8. Kg3 Qg1+ 9. Kh3 Qh1+
17 0.68 -0.08 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qxg4
11. Qd8+ Kf4 12. Qe8
17-> 0.75 -0.08 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qxg4
11. Qd8+ Kf4 12. Qe8
18 0.86 -1 1. ... c4!!
18 0.88 -0.48 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kg2 Qxg4+ 10. Qxg4+
Kxg4 11. Kf2 Kxh5 12. Kf3
18-> 0.94 -0.48 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kg2 Qxg4+ 10. Qxg4+
Kxg4 11. Kf2 Kxh5 12. Kf3 (s=4)
19 1.10 -0.48 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kh2 Qf4+ 9. Kh1 Qf3+ 10. Kg1 Qxg4+
11. Qxg4+ Kxg4 12. Kg2 Kxh5 13. Kf3
(s=3)
19-> 1.21 -0.48 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kh2 Qf4+ 9. Kh1 Qf3+ 10. Kg1 Qxg4+
11. Qxg4+ Kxg4 12. Kg2 Kxh5 13. Kf3
(s=2)
20 1.49 -1 1. ... c4!!
20 1.55 -0.78 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kh2 Kxh5 15. Kg3
20-> 1.96 -0.78 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kh2 Kxh5 15. Kg3 (s=4)
21 2.63 -1.07 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kf3 g5 (s=3)
21-> 2.99 -1.07 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kf3 g5 (s=2)
22 4.69 -1.37 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kf3 g5 16. Kg3
22-> 5.80 -1.37 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kf3 g5 16. Kg3
(s=2)
23 9.98 -1.17 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
Kg6
23-> 11.27 -1.17 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
Kg6
24 23.30 -1.38 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
g4+ 17. Kg3 Kg5
24-> 25.99 -1.38 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
g4+ 17. Kg3 Kg5
25 1:04 -1.74 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
g4+ 17. Kg3 Kg5 18. Kg2
25-> 1:09 -1.74 1. ... c4 2. bxc4 a4 3. c5 a3 4. bxa3
bxa3 5. c6 a2 6. c7 a1=Q 7. c8=Q Qf1+
8. Kg3 Qf4+ 9. Kh3 Qf3+ 10. Kh2 Qf2+
11. Kh1 Qh4+ 12. Kg1 Qxg4+ 13. Qxg4+
Kxg4 14. Kg2 Kxh5 15. Kg3 g5 16. Kf3
g4+ 17. Kg3 Kg5 18. Kg2
time=2:14 mat=1 n=703848837 fh=88% nps=5.2M
ext-> check=82.4M 1rep=4.8M mate=141K pp=0 reduce=354.1M/47.5M
predicted=0 evals=413.9M 50move=0 EGTBprobes=0 hits=0
SMP-> splits=2988 aborts=224 data=7/128 elap=2:14
-
bedouin
Re: difficult endgame test
What is the issue here? Is it a tablebase position or a test of the engines capability as there are quite a few things to be calculated.
-
Alessandro Scotti
Re: difficult endgame test
Hamsters misses but it's all in the evaluation... I followed Crafty's PV up to the last pawn capture and king move:bob wrote:I ran this with crafty... it finds this move at depth=11, .1 seconds, on my core2 duo laptop. by depth=18, .5 seconds, the score is in black's favor by 1/2 pawn.... by depth 21, it is 1.5 pawns in blacks favor, taking about 2 seconds...
[D]8/6p1/8/4p2k/3pP3/3P2K1/2P5/8 b - - 1 1
and here Crafty sees an advantage for black while Hamsters thinks this is even slightly better for white!
How do you score the passed pawn g7 in this case?
-
bob
- Posts: 20943
- Joined: Mon Feb 27, 2006 7:30 pm
- Location: Birmingham, AL
Re: difficult endgame test
I ran without tablebases myself. You can see that in the statistics where there were zero EGTB probes/hits...bedouin wrote:What is the issue here? Is it a tablebase position or a test of the engines capability as there are quite a few things to be calculated.
It is more about breaking the position open rather than locking things up..
-
bob
- Posts: 20943
- Joined: Mon Feb 27, 2006 7:30 pm
- Location: Birmingham, AL
Re: difficult endgame test
It is winning. black advances the pawn and white has to take the time to capture it. Black abandons it at the right point and penetrates to eat the white pawn chain. The only potential issue here is that in similar positions, white might be able to advance its one pawn that can move and trade the pawns, still leaving itself with time to prevent the g-pawn from queening. But in the current position, white's busted. How to defend against walking the g-pawn down. White has no moves except for the king. So black takes the opposition by moving its king, and forces white back. Even if you remove all the other pawns, black still wins this trivially.Alessandro Scotti wrote:Hamsters misses but it's all in the evaluation... I followed Crafty's PV up to the last pawn capture and king move:bob wrote:I ran this with crafty... it finds this move at depth=11, .1 seconds, on my core2 duo laptop. by depth=18, .5 seconds, the score is in black's favor by 1/2 pawn.... by depth 21, it is 1.5 pawns in blacks favor, taking about 2 seconds...
[D]8/6p1/8/4p2k/3pP3/3P2K1/2P5/8 b - - 1 1
and here Crafty sees an advantage for black while Hamsters thinks this is even slightly better for white!
How do you score the passed pawn g7 in this case?
But it is all about the concept of "distant passed pawn" here...
I'm not sure why your program would like white better. Black has a passed pawn. White has a weak pawn that can't move (backward). Black's king is in position to support its passer. Etc...
-
bedouin
Re: difficult endgame test
Trying this on some engines they get it in a fraction of a second without tablebases and I suspect that this is a well known position (could be in a testsuite of some sort) in which engines would naturally struggle. Perhaps as part of development the developers tune the endgame functions to cope with such positions. I'm impressed as the play is very deep.
-
Alessandro Scotti
Re: difficult endgame test
You are right... evaluation of this part sucks, but I've never been good in scoring passed pawns.bob wrote:I'm not sure why your program would like white better. Black has a passed pawn. White has a weak pawn that can't move (backward). Black's king is in position to support its passer. Etc...
The tables below do not even award a bonus for a passer on its home square, and the max bonus for a pawn in 7th rank is 40 to 65... maybe a bit too conservative eh?
Code: Select all
const int PassedPawnBonus_Opening[8] = {
0, 0, 5, 10, 15, 25, 40, 65
};
const int PassedPawnBonus_Endgame[8] = {
0, 0, 10, 15, 25, 40, 65, 125
};
-
Tord Romstad
- Posts: 1808
- Joined: Wed Mar 08, 2006 9:19 pm
- Location: Oslo, Norway
Re: difficult endgame test
I agree - this position tests the program's eval, not the search. Glaurung Lodz 2007, another program with a rather primitive evaluation, also struggles with this position. It eventually plays c4 at depth 21, but even after 11 minutes, 27 iterations and 1.2 billion nodes, the score is far from convincing, although the position at the end of the main line definitely is.Alessandro Scotti wrote:Hamsters misses but it's all in the evaluation...
Analysis on my MacBook Core Duo 2 GHz, 128 MB hash:
Code: Select all
18 +0.35 00:00 1116k Kf4 Kh4 c4 g5 a4 h6 axb3 hxg7 bxc2 g8=Q c1=Q Qf7+
Ke3 Qxc4 Qxc4 dxc4 d3 g6 Kxe4 g7 d2 g8=Q d1=Q
19 +0.65 00:01 1923k Kf4 Kh4 c4 g5 a4 h6 axb3 cxb3 cxd3 hxg7 d2 g8=Q
d1=Q Qf7+ Kxe4 g6 Qh1+ Kg5 Qc1+ Kf6 Qf4+ Ke6 Qg4+
Kd6 Qh5 Qd5+ Kf4
20 +0.33 00:01 2883k Kf4 Kh4 c4 g5 a4 h6 axb3 cxb3 cxd3 hxg7 d2 g8=Q
d1=Q Qf7+ Kxe4 g6 Qh1+ Kg5 Qc1+ Kf6 Qf4+ Ke6 Qg4+
Kd6 Qh5 Qb7+ Kf5 g7
21 +0.00 00:03 6233k Kf4 Kh4 c4 g5 a4 h6 axb3 cxb3 cxd3 hxg7 d2 g8=Q
d1=Q Qf7+ Kxe4 g6 Qh1+ Kg4 Qg1+ Kh3 Qe3+ Kh4 Qh6+
Kg4 Qe3 Kh4
21 +0.16 00:05 10356k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh1 Qh4+ Kg1
Qxg4+ Qxg4+ Kxg4 Kf2 Kxh5 Kg3 g5 Kf3 g4+ Kg3
22 +0.35 00:11 20804k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh1 Qh4+ Kg1
Qxg4+ Qxg4+ Kxg4 Kf2 Kxh5 Kg3 g5 Kf3 g4+ Kg3 Kg5
23 +0.45 00:26 48546k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh1 Qh4+ Kg1
Qxg4+ Qxg4+ Kxg4 Kf2 Kxh5 Kg3 g5 Kf3 g4+ Kf2 Kg5
Kg3
24 +0.65 01:01 111421k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh1 Qh4+ Kg1
Qxg4+ Qxg4+ Kxg4 Kf2 Kxh5 Kf3 g5 Ke2 g4 Kf1 Kg5
Ke2
25 +0.55 02:01 221349k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh3 Qh4+ Kg2
Qxg4+ Qxg4+ Kxg4 Kh2 Kxh5 Kh3 g5 Kg3 g4 Kg2 Kh4
Kf2 g3+ Kf3 Kh3
26 +0.76 04:07 450354k c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh3 Qh4+ Kg2
Qxg4+ Qxg4+ Kxg4 Kh2 Kxh5 Kh3 g5 Kg3 g4 Kg2 Kh4
Kf2 g3+ Kf3 Kh3 Ke2
27 +1.20 11:07 1210M c4 bxc4 a4 c5 a3 bxa3 bxa3 c6 a2 c7 a1=Q c8=Q
Qf1+ Kg3 Qf4+ Kh3 Qf3+ Kh2 Qf2+ Kh3 Qh4+ Kg2
Qxg4+ Qxg4+ Kxg4 Kh2 Kxh5 Kh3 g5 Kg3 g4 Kg2 Kh4
Kf2 g3+ Kf3 Kh3 Ke2 g2