The first time I changed some of the logical bugs of the evaluation function, but the prolonged branches were excessive, so they were still stupid on the 14th floor. . . The second revision has re-revised the twig algorithm, adds some strategies, although the search layer is shallow, but the chess is obviously stronger, and the excessive problem of the twigs is still alleviated. This version If I don't seriously, almost I will lose, but in the middle and late staggery stage, I have accumulated a thickness in some, and then use 4, 5 to kill, it is also losing. . . The problem is that it sometimes makes some boring traps, whiten waste a lot of offensive opportunities, and can't see my 5 or more, I don't know how to make the situation, I will take the array first. . . These problems are very subtle, some and the twigs are related, some and the evaluation parameters are related. If I decrease the pruning, the speed is reduced, and the adjustment of the assessment parameters is too fine. . . The current idea is to introduce annealing function, and then generate randomness in some places, I hope some help.
