From: Colin Paul Adams COLINA DEMON CO UK>
Date: 3 feb 2000
Subject: CHU: solution to A6
Again, my program disagrees with the published solutions.
It gives a Mate in 15 (Japanese count) (I have disabled usage of the
transposition table in mating mode, and the reporting errors I
mentioned for A4 promptly disappeared, and it gave the full mate in
17), compared with the published solutions of 11 and 13.
First, here is the principal continuation that the program gives:
1. Ln - 4d, GB x 4d
2. +DK x 4d, K - 2a
3. +DK x 3c - 2b, K x 2b
4. DH x 2d+, +VM x 2d (until this move, the computer was
following the alternative published solution, which gives K - 2a -
this seems to be an inferior defence)
5. DK - 4d+, K - 2a
6. Ph x 2d+, FL - 2c
7. +Ph x 2c, C - 2b
8. +DK x 2b
I then looked at the position after 4. ..., K - 2a
Sure enough, the computer finds the same Mate as published. So K - 2a
is not the best defence.
I then looked at the first published solution (starting DH x 2d+, S x
2d):
The computer failed to find a mate in the next 16 moves. So somewhere
the published defence has been defective.
After 3. +DK x 4d, I tried K - 2a, and the computer says there is no
solution.
So the primary solution is incorrect, and the alternative solution
lacks the best defence, and so mates too early.
I leave it to others to work out any sidelines.
--
Colin Paul Adams
Preston Lancashire