Geschrieben von: / Posted by: John Grant at 19 May 2000 12:31:48:
I observed a WinBoard game (on P-III 450 MHz 128 MB SD-RAM hash size of 32MB for both programs at the time control of 40 moves every 60 minutes) a victory for (as I remembered) Crux 1.2R as White over Amy as Black. The highlight of this game was an outstanding "human-like" exchange by Amy; the Black Bishop and Knight were traded for the White Pawn on f2 and Rook on f1 soon after White castled (on the King-side:O-O). In a casual rapid game I lost with White chessmen against Ian Rogers(who told my father before we played that he had the ambition to become a master or grandmaster if possible) with Black chessmen (at the chess club where he was director of play in 1970), the exactly same exchange in a very similar position was initiated by Black. Unfortunately I do not have the record of the moves in both games discussed above.
Maybe a table of more refined material values for each chessman in specific categorised positions according to number of empty chessmen and type of chessmen would help Crux and other improving programs evaluate better. For example: Queen at the beginning of the game is 9.75 pawns and doubled is 19.50 pawns when it is the only chessman on the chessboard.
With best wishes from John.