Author: Dann Corbit
Date: 17:21:48 02/04/00
Go up one level in this thread
Here is the crafty log after 15 plies have completed. Of course, the ce score
is pretty much a bunch of hooey, since it is a draw.
But I think there is no way to prove that a program *would* draw or *would*
mate, unless you play it out.
L:\>type log.014
pondering enabled.
EGTB cache memory = 6M bytes.
hash table memory = 48M bytes.
pawn hash table memory = 16M bytes.
play best book line after search.
EGTB access enabled
using tbpath=./TB
4 piece tablebase files found
PFGA: EPD record: 1
end-game phase
clearing hash tables
time surplus 0.00 time limit 116:40 (116:40)
nss depth time score variation (1)
White(1): exit
8 17.05 -4.65 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rh6 Nxc3
4. Kxc3 Rg2
8-> 17.16 -4.65 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rh6 Nxc3
4. Kxc3 Rg2
9 18.55 -4.60 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rh6 Nxc3
4. Kxc3 Rg2 5. Kd4
9-> 18.59 -4.60 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rh6 Nxc3
4. Kxc3 Rg2 5. Kd4
10 24.78 -4.70 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Rh6 Nba4
4. c4 Nb2 5. Kc3 Nxc4 6. Kxc4
10-> 25.01 -4.70 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Rh6 Nba4
4. c4 Nb2 5. Kc3 Nxc4 6. Kxc4
11 26.45 -- 1. Ke3
11 46.88 -5.13 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Kc1 Nba4
4. Rg3 Ne4 5. Rd3+ Kc6 6. c4 Kc5 7.
Ra3 Kxc4 8. Rxa4+
11-> 48.90 -5.13 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Kc1 Nba4
4. Rg3 Ne4 5. Rd3+ Kc6 6. c4 Kc5 7.
Ra3 Kxc4 8. Rxa4+
12 50.68 -- 1. Ke3
12 1:33 -5.92 1. Ke3 Nc5+ 2. Kd4 Re4+ 3. Kd5 Nf4+
4. Kxc5 Nxg6 5. Kd5 Re5+ 6. Kd4 Rb5
7. b4 Ne5 8. Ke4 Kc6 9. Ke3 Rd5
12-> 1:35 -5.92 1. Ke3 Nc5+ 2. Kd4 Re4+ 3. Kd5 Nf4+
4. Kxc5 Nxg6 5. Kd5 Re5+ 6. Kd4 Rb5
7. b4 Ne5 8. Ke4 Kc6 9. Ke3 Rd5
13 3:35 -- 1. Ke3
13 4:53 -6.52 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rxg5 Nxg5
4. Kc5 Ne6+ 5. Kb6 Kd6 6. Kb5 Nc7+
7. Kb4 Kd5 8. Kb3 Nc4
13-> 4:56 -6.52 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rxg5 Nxg5
4. Kc5 Ne6+ 5. Kb6 Kd6 6. Kb5 Nc7+
7. Kb4 Kd5 8. Kb3 Nc4
14 14:33 -6.56 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rxg5 Nxg5
4. Kc5 Ne6+ 5. Kb6 Kd6 6. Kb5 Nf4 7.
Kb4 Kd5 8. Kb5 Nc4
14-> 14:39 -6.56 1. Ke3 Nxb2 2. Kd4 Rg5 3. Rxg5 Nxg5
4. Kc5 Ne6+ 5. Kb6 Kd6 6. Kb5 Nf4 7.
Kb4 Kd5 8. Kb5 Nc4
15 16:34 -- 1. Ke3
15 72:24 -7.02 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Rg4 Ne4+
4. Kc2 Na4 5. c4 Nd6 6. Kb3 Nc5+ 7.
Kc3 Nce4+ 8. Kb3 Rc5 9. Rxe4 Nxe4 10.
Kb4
15-> 73:33 -7.02 1. Ke3 Nc5+ 2. Kd2 Nxb2 3. Rg4 Ne4+
4. Kc2 Na4 5. c4 Nd6 6. Kb3 Nc5+ 7.
Kc3 Nce4+ 8. Kb3 Rc5 9. Rxe4 Nxe4 10.
Kb4
Two knights are not going to mate very often, if both programs have tablebases.
I think a KNNNk tablebase might be fun, though. Playing OTB, give me two
bishops or a rook, at least or I'll be wandering in circles. At least I can
memorize those algorithms. I am pretty sure I could not execute a mate with two
knights, except by accident.
Playing against a chess program, I would try to manuver an exchange of both
knights for the rook. It might seem good to the chess program since the points
add up, but it would be an awful trade.
;-)
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.