Author: Tim Foden
Date: 06:18:32 04/17/04
Go up one level in this thread
On April 17, 2004 at 08:16:21, Joachim Rang wrote:
>This position is an easy mate in 4:
>
>[D]6Q1/8/8/7k/8/8/3p1pp1/3Kbrrb w - - 0 1
>
>Certain engines however have problems to see the mate and return instead a draw
>score. So ie Crafty.
>
>I think it is related to nullmove, since the solution involves a zugzwang for
>the black side and since it's so many material on board most engines will use
>nullmove in this position. However I thought Nullmove will produce a reduction
>rathern than a complete cutt-off and hence the solution should be found a few
>plies later. Or did I understand the mechanism of Nullmove wrong?
You're pretty much right, but you missed a bit of the logic. The real premise
behind NULL MOVE is that it is ALWAYS worse to make NO MOVE than to actually
make a move.
As long as this premise holds true in the positions searched, then null move
will work as you describe... as a depth reduction.
However, if there is a zugzwang in the search, then one side will find that in
fact, NO MOVE is the best move, and none if the actual moves will be seen. Thus
the engine can be completely blind to the problem, even to the maximum search
depth.
>I know that this position is not of any interest if it comes to playing
>strength, but it is rather funny to watch certain engines reach depth 20 quickly
>while being unable to spot the mate.
:) Current released GLC has this problem too.
GLC 3.00:
21 5.198 +0.000 8839954 Qg3 Kh6 2. Qg8 Kh5 3. Qg3
22 5.438 +0.000 9232834 Qg3 Kh6 2. Qg8 Kh5 3. Qg3
My Graz version is better here (it has more restrictions on when the NULL MOVE
heuristic can be used)...
GLC 3.00.3.4 (Graz):
5 0.030 -14.304 11130 Qg3 Kh6 2. Qg4 Kh7 3. Qd4
6 0.030 -13.904 13076 Qg3 {++} Kh6 2. Qg8 Kh5 {ht}
6 0.030 +0.000 13155 Qg3 Kh6 2. Qg8 Kh5 3. Qg3
6 0.040 +0.000 26206 Qg3 Kh6 2. Qg8 Kh5 3. Qg3
7 0.040 +0.000 32584 Qg3 Kh6 2. Qg8 Kh5 3. Qg3
7 0.060 +0.400 54224 Qg7 {++} Kh4 2. Qg6 Kh3 3. Qg5 Kh2 {ht}
7 0.060 +Mate04 54683 Qg7 Kh4 2. Qg6 Kh3 3. Qg5 Kh2 4. Qh4#
7 0.060 +Mate04 57266 Qg7 Kh4 2. Qg6 Kh3 3. Qg5 Kh2 4. Qh4#
Cheers, Tim.
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.