If that is an "early" resignation, I would hate to see a "late" one. It only resigned when mate was inevitable. It was down a rook and two pawns for 25 moves with zero compensation, but it still did not resign. In one game I set the program to resign "at the first sign of trouble". To add insult to injury, there are other, irritating flaws in the program. In the endgame where it cannot think for more than a few seconds, I would estimate it only plays around 1400 ELO - pathetic for a product that is advertised to play at 2320 ELO and which has some pretty decent hardware to run on. But because it moves extremely quickly, it's playing strength is far below this. I would go as far as to say that if it actually used the correct amount of time as per the level settings chosen by the user, then its playing strength might well approach 2320 ELO. There is no denying that the actual playing engine behind this product is extremely competant. This is why I say that Fritz DS is a missed opportunity. Even if it's excellent middle game tactical ability confers some advantage going into the endgame, the endgame is so poor that any advantage often rapidly evaporates. Quite simply, Fritz does not give itself anywhere near enough time to come up with a competent move. In some games against my Mephisto Miami dedicated computer (which is known itself to have a weak endgame), Miami could simply pick off Fritz's pawns almost at will. Given the program does not appear to utilise hash tables, this makes it's endgame incredibly weak indeed. Even when set to make one move within a day, Fritz still moves literally within seconds when in an endgame situation with a minor piece and pawns on the board. In the endgame, things get far worse again. Anyone who is knowledgable about chess computers will tell you that using a tiny fraction of the playing time available will severely weaken the playing strength of a chess program, simply because the program is unable to examine as many moves as it would otherwise be able to.Īnd this is the crux of Fritz's problems on the DS platform. At the default Blitz levels - even when set to 2320 ELO, the program moves rediculously quickly and barely uses any time at all. It is all but impossible to make Fritz think for more than around 80 seconds in the middlegame (except in extremely complicated situations), even if the playing level is set to one move in 23 hours and 59 seconds! And if the playing level is set to something like 40 moves in 2 hours, then Fritz DS will move extremely fast - within seconds infact.
The biggest single problem with this version of Fritz is that the level algorythms are inexcusably buggy. Unfortunately, the performance of this product on paper is completely at odds with it's performance in reality. This is an extremely high playing level and only 80 points below the rating required to achieve a human International Master title. Even when one actually plays against Fritz on the Nintendo DS, the quoted rating at the top level is 2320 ELO. Indeed, the official advertising for Fritz DS makes mention of the historical feats of Fritz and strongly implies that Fritz himself has finally come to the rescue of Nintendo console users. Afterall, the Nintendo DS console has hardware comparable to that contained in some dedicated chess computers and those machines play around the 2400 ELO level.
One would have hoped that this version of Fritz for Nintendo DS might demonstrate at least some of the incredible playing strength of it's PC based namesake.