Author: Gareth McCaughan
Date: 06:52:19 10/28/01
Go up one level in this thread
Roy Eassa wrote:
> There were several K+P positions during the game (or in the notes) that I
> consider unclear. Can anybody provide a clear view on any of these?:
...
> [D] 8/8/1k6/2pPpppp/2P5/5P2/4K1PP/8 w - - 0 42
I set Crafty looking at this for a good long time
on my machine (Athlon/1GHz, fairly modest amount of
hash). I find its analysis implausible. Some extracts
follow.
24 145:52 1.75 1. Kf2 Kc7 2. g4 hxg4 3. fxg4 f4 4.
h4 gxh4 5. g5 h3 6. g6 e4 7. g7 e3+
8. Ke2 h2 9. g8=Q h1=Q 10. Qf7+ Kc8
11. Qxf4 Qc1 12. Qf8+ Kd7 13. Qf7+
Kd8 14. Qf6+ Ke8 15. Qg6+ Ke7 16. Qe6+
Kf8 17. Qd6+ Kf7 18. Qxc5
Surely the final position here is dead drawn. (Crafty
thinks so too when given that position for a few minutes.
Or indeed a few seconds.)
26-> 1346:24 2.20 1. Kf1 Kc7 2. g4 hxg4 3. fxg4 fxg4
4. Ke2 Kd6 5. Ke3 Ke7 6. Ke4 Kf6 7.
d6 Ke6 8. d7 Kxd7 9. Kxe5 Ke7 10. Kd5
Kf6 11. Kxc5 Ke5 12. Kb6 g3 13. hxg3
g4 14. c5 Kd5
Surely 3...fxg4 is a dreadful move and 3...f4 is much
stronger. Perhaps Black is avoiding 3...f4 because of
lines like the one above where Crafty overestimates
White's chances in the two-queen position.
Now, I'm a patzer, and Crafty is supposed to be good
at K+P endings, so maybe I'm wrong about these...
> [D] 8/8/1k1P4/1Pp3p1/2K1p3/7P/7P/8 w - - 0 45
After about an hour, with a much larger hash table than
for the previous position, Crafty is convinced that this
is a draw. The variations it displays aren't much help.
Here's its complete output.
nss depth time score variation (1)
10 0.22 -- 1. Kc3
10 0.36 1.26 1. Kc3 Kb7 2. Kd2 c4 3. Kd1 Kb6 4.
Kc2 e3 5. Kc3 Kb7
10-> 0.44 1.26 1. Kc3 Kb7 2. Kd2 c4 3. Kd1 Kb6 4.
Kc2 e3 5. Kc3 Kb7
11 0.60 1.22 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 e3 5. Kd3 c4+ 6. Kxe3 c3
11-> 0.75 1.22 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 e3 5. Kd3 c4+ 6. Kxe3 c3
12 0.96 1.04 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 c4 5. Kd2 Kb7 6. d7 Kc7 7. Kc3
Kxd7 8. Kxc4
12-> 1.23 1.04 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 c4 5. Kd2 Kb7 6. d7 Kc7 7. Kc3
Kxd7 8. Kxc4
13 1.58 1.17 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 c4 5. Kd1 Kb7 6. Kd2 Kb6 7. d7
Kc7
13-> 1.97 1.17 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc2 c4 5. Kd1 Kb7 6. Kd2 Kb6 7. d7
Kc7
14 2.45 -- 1. Kc3
14 4.07 0.58 1. Kc3 Kb7 2. Kd2 c4 3. Ke3 c3 4. Ke2
Kb6 5. Ke1 e3 6. Kd1 Kb7 7. Ke2 c2
8. d7 Kc7
14 11.29 0.75 1. h4 gxh4 2. h3 Kb7 3. Kxc5 e3 4.
b6 e2 5. d7 e1=Q 6. d8=Q Qc3+ 7. Kd6
Qd2+ 8. Ke7 Qe3+ 9. Kf7 Qf3+ 10. Qf6
Qxh3
14-> 11.30 0.75 1. h4 gxh4 2. h3 Kb7 3. Kxc5 e3 4.
b6 e2 5. d7 e1=Q 6. d8=Q Qc3+ 7. Kd6
Qd2+ 8. Ke7 Qe3+ 9. Kf7 Qf3+ 10. Qf6
Qxh3
15 12.30 -- 1. h4
15 16.83 0.53 1. h4 gxh4 2. Kc3 h3 3. Kc4 Kb7 4.
Kxc5 e3 5. b6 e2 6. d7 e1=Q 7. d8=Q
Qc3+ 8. Kd6 Qd4+ 9. Ke7 Qe5+ 10. Kd7
Qxh2
15 20.03 0.94 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc4 <HT>
15-> 20.44 0.94 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc4 <HT>
16 21.20 -- 1. Kc3
16 25.05 0.53 1. Kc3 Kb7 2. h4 gxh4 3. Kc4 Kb6 <HT>
16-> 30.84 0.53 1. Kc3 Kb7 2. h4 gxh4 3. Kc4 Kb6 <HT>
17 39.20 ++ 1. Kc3!!
17-> 57:05 0.92 1. Kc3 Kb7 2. h4 gxh4 3. Kc4 Kb6 <HT>
18 57:33 0.75 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc4 <HT>
18-> 57:54 0.75 1. Kc3 Kb7 2. h4 gxh4 3. h3 Kb6 4.
Kc4 <HT>
19 58:19 -- 1. Kc3
19 59:59 0.00 1. Kc3 Kb7 2. Kc4 Kb6
19-> 62:30 0.00 1. Kc3 Kb7 2. Kc4 Kb6
20 63:21 0.00 1. Kc3 Kb7 2. Kc4 Kb6
20-> 67:21 0.00 1. Kc3 Kb7 2. Kc4 Kb6
21 69:15 0.00 1. Kc3 Kb7 2. Kc4 Kb6
21-> 75:24 0.00 1. Kc3 Kb7 2. Kc4 Kb6
Presumably it's discovered that pushing the b, c, d, g, h
pawns always makes matters strictly worse for the player
who does it, and that there's nothing else for either
player to do. :-)
I also set Crafty on
[D] 8/1k6/3P4/1Pp3p1/2K1p2p/7P/6P1/8 w - - 0 45
for about an hour and once again I am highly unconvinced
by the results. The latest PV produces another of those
two-queens positions:
21-> 41:12 1.85 1. Kxc5 e3 2. b6 e2 3. d7 e1=Q 4. d8=Q
Qf2+ 5. Kd6 g4 6. Qc7+ Ka6 7. hxg4
Qf6+ 8. Kd5 Qg5+ 9. Ke6 Qg6+ 10. Ke7
Qg7+ 11. Ke6 Qg6+ <HT>
and the position at the end looks very drawn. Crafty
agrees when given that position to think about for a
few minutes: eval of 0.00. It looks like it's found a
perpetual check.
--
g
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.