From: Steve Evans NETSPACE NET AU> Date: 9 jan 1999 Subject: Chu shogi mating problems I have a dual purpose in writing this message. The first is to give a plug for a new piece of commercial software that among many other things, plays shogi and the 5x5 variant mini-shogi. Zillions-of-Games is billed as the first universal gaming system, and seems to live up to this tag. The commercial package combines a powerful game playing engine, with capacity for playing board games over the internet, and comes with 292(!) different games (including shogi). The other attractive feature about the package is that it provides a reasonably simple programming language that allows users to write their own games. You can find out more about Zillions of Games from their home page at: http://www.zillions-of-games.com This brings me to my second reason for writing. I recently decided to write a chu shogi driver for the Zillions engine, and have now finished this exercise. The chu rules file I've written should shortly be available for download from the Zillions web site (it needs the Zillions package to run). I decided to use Zillions to implement a chu shogi program for a few reasons: 1) to see how easy it was to write a complex game for the Zillions engine; 2) to make freely available a chu playing program that plays a much stronger game than any others currently about (including my own Shogi Variants program); & 3) because I wanted to adapt it to solve some of the 120+ ancient chu shogi mating problems given in George Hodges' Middle Shogi Manual for which no solutions have ever been recorded. Yesterday evening I finished my mating problem adaptation of the program and decided to test it on one of the problems for which the solution is known. For the exercise I chose a problem by Ito Sokan I from his collection published in 1678. The problem (in modified Forsyth Notation) is as follows: 3, S, 1, PH, K, G, 4 / 4,+dk, 1, DE, 2, S, 2 / 2, +DK, ln, 5, FL, 2 / 5, C, 2, P, 3 / 3, P, 2, +dh, ky, 4 / 4, +VM, 7 / 12 / 2, +vm, 9 / 6, FK, 5 / 12 / 12 / 12 / This is problem B1 in the Middle Shogi Manual (for those who own a copy). The solutions given in the manual are: 1. +VM-10e, K-5b (if 1. ..., +DKx10e; 2. Lnx7a, DEx7a; 3. +DK-6d, +VMx6d; 4. +DH-6c, +VMx6c; 5. Kyx5c+ mate) 2. +DHx3b, FLx3b; 3. Kyx4d+, K-4a; 4 +Ky-2b, K-5b; 5. Lnx7d mate. The program found the first of these two solutions in about 3 minutes, so when I've tested it a bit more I hope to tackle some of the unsolved problems. While I suspect that problems with solutions deeper than 6 or 7 moves will be beyond it, I'm hopeful that some of the unsolved problems will have shorter solutions than this. I can also use the program to verify and test some of the published solutions to these old mating problems, and as a tool to assist in solving some of the longer problems. If I have any success I'll announce it on Shogi-L. Regards, Steve Evans