Kretland

Ultra Gomoku 2.0 - description

Version 2.0 of the program is a development of the first DOS version of the game. It contains extensive pieces written in Assembler under DOS, which is why it is a 16-bit Windows 3.1 application. This, however, has no impact upon the power of the game. The only limitation is that the names of saved and restored files recording the course of the rounds must be in the DOS format, that is 8.3 (8 characters of the name and 3 of the extension).

The game can be played on a 15x15 or 19x19 board. You can undo your moves without limitation. There is also the option of setting up the initial situation, which makes Ultra Gomoku 2.0 a perfect tool for problem solving. The rounds can be saved and restored, of course. The program has three levels of difficulty. At your request it can prompt you, analysing the game at the currently running level.

The program plays standard gomoku. There are other variations of the game apart from this, Pro gomoku and Swap gomoku, however, they are not available in the program. The most significant differences concern opening moves and the overline (that is, 6 or more stones in a row). Whereas the opening moves do not influence solving problems in any way, the overline, unfortunately, does. Namely, while in standard gomoku, the game played by this program, an overline is recognized as a win, it is not so in the other two variations of the game. This is why this program, following the rules of standard gomoku, will not correctly solve problems with overline for pro or swap gomoku. Consequently, tasks involving overline were omitted when solving problems from Polish Association of Gomoku, Renju and Pente's website. However, there were only 7 of them ( out of the total of 85).

After leaving out the tasks with overline it seemed that the program could not cope with four of the other problems. However, when they were verified by their author, it turned out that the program was right: three of them (task 19, 51 and 62) could not be solved in the way suggested and had to be changed! In the fourth one an error appeared while transferring it to the Ultra Gomoku program. When corrected, all the problems were solved.

Thus the test results prove the program to be 100% effective in problem solving. Obviously, this gives you no guarantee that all tasks will always be solved correctly, but testing the program on a significant number of 78 tasks shows that it is a good tool for solving gomoku problems.

It is worth mentioning that some of the solutions proposed by the program seem more interesting than the ones suggested by the author of the tasks. However, it has to be admitted that sometimes (fortunately, rarely) the program makes unnecessary moves. Almost all the tasks were solved at level 2.

Thanks to extremely fast recurrent algorithm written in Assembler, Ultra Gomoku 2.0 spots 100% of the VCF (Victory by Continuous Fours) situations, no matter how many moves are needed to accomplish such a victory. This is well shown by the examples of the solved tasks. The program promptly and easily finds VCF solutions which require sometimes 20-30 (or even more!) moves. Of course, the algorithm also solves the "cut" problem, that is, a situation in which the player blocking a four set up by his opponent sets up a four of his own that needs to be blocked. Such a four breaks the victorious VCF sequence, but the "cut" situations are effectively avoided by the program. No VCF combination that Ultra Gomoku 2.0 is unable to find has been reported so far. It has to be admitted, though, that it may theoretically appear one day.

The VCT problem (Victory by Continuous Threats) has, as most problems in Ultra Gomoku 2.0, also been solved by algorithms written in Assembler. In this case, however, it is not a final solution yet, as it is with VCF.