The entire program and its style of play is reminiscent of Philidor’s Intelligent Chess, although there is a greater emphasis on pawn moves. The evaluation function is evidently very sophisticated, but it takes a long time to search at the lowest levels. It seems that subsequent levels are only sketchily evaluated, and the program rarely changes its mind about a move found at the lowest level. As a result, PetChess falls into simple traps at these levels – up to one minute, which covers half of the available levels – because of its inability to see far ahead. At higher levels, it avoids the traps, but tends to exchange everything it can. Like so many programs, it likes to exchange its bishops for knights early in the game, and it also prefers to double its pawns rather than move its queen before castling – I detect Levy’s hand. PetChess shows some evidence of strategic thinking in moving pawn majorities against minorities or in moving major pieces into attack positions against the enemy king. This, again, shows the thought that has been put into the evaluation function. The end-game play is, however, rather weak, and the king seems to be a little passive.