Darren Cook wrote: > All except joseki-knowledge is board-size independent.
Maybe human player's adapt to different board sizes without even noticing. But if you try to model strategy with algorithms it is totally board size dependent. The extreme case is 5x5 where black 3,3 claims the four corners. The relative size of the corner, the sides and the center is crucial. Move timing, distances between stones, everything. I like use the "two day argument" because I believe a top level computer go program should have "two phases" (But the first, of course will be much longer than 50 moves.) 2nd. When all local fights can be searched. If is a matter of understanding their conditions and interactions. The program is an endgame solver. It plays "kami no itte" for the last ¿m? moves. Complexity increases with board size, but in a known way. 1st. All the rest of the game. In this part there is no certitude. The program uses stochastical evaluation (UCT) together with knowledge to "navigate in the fog" expecting to be ahead when the endgame solver "finds the harbor". *All* the knowledge used in the 1st step is board size dependent. UCT is not, but UCT's efficiency (and variance) is. Jacques. _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
