Author: Michael Fuhrmann
Date: 15:30:03 04/23/00
Go up one level in this thread
On April 23, 2000 at 18:13:42, Robert Hyatt wrote:
>On April 23, 2000 at 17:15:52, Michael Fuhrmann wrote:
>
>>Why would a program ever underpromote to a rook? Saw Crafty do this recently.
>>(In this particular case, it had no impact on the outcome of the game.)
>
>
>1. it is necessary at times. IE if you promote to queen, you stalemate your
>opponent. if you promote to rook, you can still win without stalemating him.
>
Of course. Didn't think of that.
>2. In the case of chess engines, it is pretty common to see this. The most
>common reason is that the =R is not a check, when the =Q is a check, or the
>rook allows fewer checks later in the tree. So by promoting to a rook, it
>avoids some tactic that it really can't avoid... IE this is a horizon effect
>situation..
Not sure I follow that. But I guess that was the case. Here's the relevant bit
cut-and-pasted from the log file, in case anyone's interested. The
underpromotion occurs on the a file. Neither =Q or =R leads to check.
White(49): Bb7
time used: 7.54
+---+---+---+---+---+---+---+---+
8 | | | | *R| | | | |
+---+---+---+---+---+---+---+---+
7 | P | B | *B| | | | *K| |
+---+---+---+---+---+---+---+---+
6 | | | P | | | *P| | |
+---+---+---+---+---+---+---+---+
5 | | | B | *P| | | *P| P |
+---+---+---+---+---+---+---+---+
4 | | | | | | | | |
+---+---+---+---+---+---+---+---+
3 | | | | | | K | P | |
+---+---+---+---+---+---+---+---+
2 | | | | | | P | | |
+---+---+---+---+---+---+---+---+
1 | | | | | | | | |
+---+---+---+---+---+---+---+---+
a b c d e f g h
Black(49): time 23196
time remaining: 3:51 (crafty).
Black(49): otim 19997
time remaining: 3:19 (opponent).
Black(49): f6f5
time used: 6.42
time surplus 0.00 time limit 7.48 (44.88)
nss depth time score variation (1)
9 0.99 6.97 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kxh5 54. Kxd5 Kg5 55. Bb7
9-> 1.24 6.97 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kxh5 54. Kxd5 Kg5 55. Bb7
10 1.74 6.93 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kxh5 54. Kxd5 Kg6 55. Bb7
f4
10-> 2.47 6.93 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kxh5 54. Kxd5 Kg6 55. Bb7
f4
11 4.24 7.02 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kg5 54. Kxd5 Kxh5 55. Bb7
Kg5 56. Be7+ Kg6 <HT>
11-> 4.76 7.02 50. a8=R Rxa8 51. Bxa8 g4+ 52. Ke3
Kh6 53. Kd4 Kg5 54. Kxd5 Kxh5 55. Bb7
Kg5 56. Be7+ Kg6 <HT>
time=7.60 cpu=100% mat=0 n=1595624 fh=92% nps=209950
ext-> checks=67434 recaps=8983 pawns=10271 1rep=704 thrt:12
predicted=17 nodes=1595624 evals=319839
endgame tablebase-> probes done=0 successful=0
White(50): a8=R
time used: 7.60
+---+---+---+---+---+---+---+---+
8 | R | | | *R| | | | |
+---+---+---+---+---+---+---+---+
7 | | B | *B| | | | *K| |
+---+---+---+---+---+---+---+---+
6 | | | P | | | | | |
+---+---+---+---+---+---+---+---+
5 | | | B | *P| | *P| *P| P |
+---+---+---+---+---+---+---+---+
4 | | | | | | | | |
+---+---+---+---+---+---+---+---+
3 | | | | | | K | P | |
+---+---+---+---+---+---+---+---+
2 | | | | | | P | | |
+---+---+---+---+---+---+---+---+
1 | | | | | | | | |
+---+---+---+---+---+---+---+---+
a b c d e f g h
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.