On Thu, 2007-12-20 at 13:36 -0500, Don Dailey wrote: > The only way this might help is in the opening or in very nearly > symmetrical positions and this is really rare. The possible slight > benefit would be canceled by even a very small slowdown. > > It would be useful on small boards as an opening book however where > exact positions (or hashes) are stored.
And that was *exactly* what the OP was asking. For fuseki/joseki collections, is is easy to transpose the board in such a way that most of the stones are in one corner. Preferably the lower left corner, since the X/Y-coordinates are smallest there. A simple function , such as sum(black)+ 361*sum(white), where 'black' and 'white' are the lineair coordinates (eg x+19*y) can be used for finding the configuration where _most_ of the stones are in the preferred corner. (Other functions can be used.) A better (but more difficult to implement) method would be transpose the board in such a way, that the stone that is closest to any edge is located to a particular corner/edge (eg left bottom) . proceed to the second closest until the tie-bereaker is found. If not: the position *is* symmetric) This method *is* costly, but that may be no problem when building a database. When using is, one has to check against all the 8 or 16 configurations, which will differ in zobrist-hash (most of the time ;-). Normally, the board will need to be flipped only a few times, using this method; once one of the four corners becomes the crowded spot, it will be the lower left, and stay that way for very long. NB: The 'order by mass' method is used in stereo-chemistry: the four neighboring atoms of a tetraeder are ordered by atomic mass, and the compound can then be classified as L- or R-isomer. HTH, AvK _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
