Author: Dieter Buerssner
Date: 09:59:42 12/29/00
Go up one level in this thread
On December 29, 2000 at 12:00:41, Uri Blass wrote:
>This position is similiar to the position that crafty had no problem to solve
>
>[Refer to original post for board displ.]R7/8/6k1/4n3/5p2/5P2/5r1p/7K b - - 0 1
>
>The test is to avoid Nxf3 that is leading to a draw.
Yace, this time with EGTBs disabled, need 39 seconds to avoid the move,
but almost double of this time to come up with a PV. Note that the moves,
that have a 'H' appended are retrieved from the transposition tables at
display time, and not at search time. So they can posssibly not be trusted,
because they may have been updated in between.
nodes time score depth PV {material balance at end of PV}
8901907 34.796 6.59 12. 1...Nxf3 2.Ra6+ Kf5 3.Rf6+ Ke5 4.Rf5+ Ke6
5.Rf6+ Kd5 6.Rf5+ Kd6 7.Rf6+ Kc5 8.Rf5+ Kc6
9.Rf6+ Kd7 10.Rd6+ Kc7 11.Rc6+ Kb7 12.Rb6+ Ka7
13.Ra6+ Kb8 {500}
9232519 36.181 6.19 13-- 1...Nxf3 2.Ra6+ Kf5 3.Rf6+ Ke5 4.Re6+ Kd4
5.Rd6+ Kc3 6.Rc6+ Kb4 7.Rb6+ Kc4 8.Rc6+ Kb5H
9.Rb6+H Kc4H 10.Rc6+H Kb5H 11.Rb6+H Kc4HR {500}
9768772 38.257 1.36 13t 1...Nxf3 2.Ra6+ Kf5 3.Rf6+ Ke5 4.Re6+ Kd4
5.Rd6+ Kc3 6.Rc6+ Kb4 7.Rb6+ Kc4 8.Rc6+ Kd3
9.Rd6+ Ke4 10.Re6+ Ne5 11.Rxe5+ Kf3 12.Re3+ Kg4
13.Re2 Rf1+ 14.Kxh2 f3 15.Re4+ Kg5 16.Re8H f2H
17.Kg2H Kf6H 18.Kxf1H Kg5H {HT}
9828483 38.541 1.37 13t+ 1...Rxf3 2.Kxh2
20136873 1:20.1 5.52 13t 1...Rxf3 2.Ra6+ Kf5 3.Kxh2 Ng4+ 4.Kg1 Rf2
5.Ra5+ Kf6 6.Ra6+ Kg5 7.Ra5+ Kh4 8.Ra8 f3
9.Rh8+ Kg5 10.Rc8 Rg2+ 11.Kf1 Ne5 {420}
>Another test is to see the draw after Nxf3.
After a bit more than one minute, I get draw score.
9774919 39.778 -5.06 16++ 2.Rg8+ Kf5 3.Rg5+ Ke6 4.Re5+ Kd7 5.Rd5+ Ke7
6.Rd7+ Kf6 7.Rd6+ Kg5 8.Rd5+ Kh6 9.Rh5+ Kg7
10.Rg5+ Kh6 11.Rg6+ Kh5 12.Rg5+ Kh4 13.Rh5+ Kg4
14.Rg5+H Kh3H 15.Rh5+H Nh4H 16.Rxh4+H Kg3H
17.Rxf4H Rf1+H 18.Rxf1H Kh4H {HT}
13288328 52.972 -1.30 16t 2.Rg8+ Kf5 3.Rg5+ Ke6 4.Re5+ Kd7 5.Rd5+ Ke7
6.Rd7+ Kf6 7.Rd6+ Kf5 8.Rd5+ Ne5 9.Rxe5+ Kg4
10.Rg5+ Kf3 11.Rg3+ Ke4 12.Re3+ Kd4 13.Rd3+ Kc5
14.Rd5+ Kb4 15.Rd4+ Kb5 16.Rd2 Rf3 17.Rb2+ Kc5
18.Rc2+ Kd6H 19.Rd2+H Ke5H 20.Kxh2H {-80}
13313702 53.103 -1.30 16. 2.Rg8+ Kf5 3.Rg5+ Ke6 4.Re5+ Kd7 5.Rd5+ Ke7
6.Rd7+ Kf6 7.Rd6+ Kf5 8.Rd5+ Ne5 9.Rxe5+ Kg4
10.Rg5+ Kf3 11.Rg3+ Ke4 12.Re3+ Kd4 13.Rd3+ Kc5
14.Rd5+ Kb4 15.Rd4+ Kb5 16.Rd2 Rf3 17.Rb2+ Kc5
18.Rc2+ Kd6 19.Rd2+ Ke5 20.Kxh2 {-80}
15500082 1:00.7 -0.90 17++ 2.Rg8+ Kf5 3.Rg5+ Ke6 4.Re5+ Kd7 5.Rd5+ Ke7
6.Rd7+ Kf6 7.Rf7+ Ke5 8.Re7+ Kd5 9.Re5+ Kc4
10.Rc5+ Kb3 11.Rc3+ Kb4 12.Rc4+ Ka5 13.Rc5+ Ka4
14.Rc4+ Kb5 15.Rc5+ Ka6 16.Rc6+ Ka7 17.Ra6+ Kb7
18.Rb6+ Kc7 19.Rc6+ Kb8H 20.Rb6+H Ka8H 21.Ra6+H
Kb7H 22.Rb6+H Kc7H 23.Rc6+H Kb8H 24.Rb6+H Ka8H
25.Ra6+H Kb7HR {-500}
16661904 1:05.0 0.00 17t 2.Rg8+ Kf5 3.Rg5+ Ke6 4.Re5+ Kd7 5.Rd5+ Ke7
6.Re5+ Kd6 7.Rd5+ Kc7 8.Rc5+ Kd8 9.Rd5+ Ke8
10.Re5+ Kf7 11.Re7+ Kg6 12.Re6+ Kh5 13.Re5+
Nxe5 {-1000}
>Did somebody try to do a program not for playing games but only to find if there
>is a perpetual check?
I do not use your idea, but I do some extensions, that should help to
detect perpetual checks (by search). If this makes the program stronger is yet
another question ...
Also, in my experience, the depth, when detecting the perpetual check,
can depend very much on hash table size. As you know, the path to the position
is not hashed. I try to be as careful when I retrieve entries from the
hash. I.e. I allways check, if the move yields in a repeted position, and won't
allow the move, if this happens, and the score is not a draw score. But
this is only a small help, and the general problem seems to be almost
unsolvable, without giving up almost all of the efficiency of the hash.
Nevertheless, in some positions I checked, it did help.
-- Dieter
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.