quotMajorquot Deep Fritz 10 Bug


Back to forum

Ron Keyston    (2007-01-29)
"Major" Deep Fritz 10 Bug

I've confirmed this problem on two different computers with completely different hardware and different operating systems. I've also sent the problem off to Chessbase, but have not yet gotten much of a response. If anyone else has Deep Fritz 10, would you mind giving this a try and reporting back with your results? Also, if anyone has the non-Deep version of Fritz 10, I'd be interested in knowing if it is also affected by this problem.

Input a game into Deep Fritz 10 and get to a point in the game where it is possible for black to castle long. Now save the game into a database, close the game and then open it back up from the database. If you either turn on infinite analysis, or just try to make the move, black is not able to castle long...Fritz assumes that it is an illegal move.

Furthermore, if you castle long BEFORE saving the game into the database, then save it and re-open it, then go to the position after black has castled queenside and turn on infinite analysis, the analysis is "messed up." Either the analysis text is invisible, or it reports impossible lines, or the evaluation score is very obviously wrong. This should be enough info for anyone to give the test a try, but if you want some specific examples, please let me know.


Miguel Pires    (2007-01-29 17:42:25)

Can you give some examples? And another thing, waht Hash yable you use? And what is the maximum HT you can load? Best Regard's Miguel Pires

Ron Keyston    (2007-01-29 19:04:14)
Specifics and Examples

OK, one PC is a 3.2GHz P4 w/1GB RAM running XP Pro. 256MB Hashtables with an ~800MB Maximum possible. The other PC is an Athlon X2 4600+ w/2GB RAM running Vista Ultimate (RTM.) It has 1GB Hashtables with ~1.5GB Maximum Possible. I can pretty much guarantee that it is not a hashtable issue though as the problem is with the legality of a move and only arises after saving into a database, closing the game and then re-opening it from the database. Also, if I open the same saved game from the database into Fritz 9, castling long as black is perfectly OK.

As for some examples, I will give the same five games that I sent to Chessbase. Some of these are contrived examples, whereas some are from my games here at FICGS:

1.e4 d5 2.exd5 Qxd5 3.Nc3 Qd6 4.Nf3 Nf6 5.Bc4 Bg4 6.O-O Nc6 7.d4

1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.Bg5 dxe4 5.Nxe4 Be7 6.Bxf6 gxf6 7.Nf3 b6 8.Bc4 Bb7 9.Qe2 Nd7 10.O-O-O c6 11.Rhe1 Qc7 12.g3

1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3 Bxc3+ 6.bxc3 c4 7.Qg4 g6 8.Nf3 Qa5 9.Bd2 Nh6 10.Qh3 Nf5 11.g4 Nxd4 12.cxd4 Qb6 13.Bg2 Nc6 14.Qh6 Nxd4 15.O-O Bd7 16.Bg5 Ba4 17.Nxd4 Qxd4 18.Qg7 Rf8 19.Be3 Qxg4 20.Bc5

1.e4 d5 2.exd5 Qxd5 3.Nc3 Qd6 4.d4 Nf6 5.Nf3 a6 6.Be3 Nc6 7.Qd2 Bf5 8.Bd3 Bg4 9.Be2 e6 10.O-O-O

1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 c5 5.a3 Bxc3+ 6.bxc3 c4 7.Qg4 g6 8.Nf3 Qa5 9.Bd2 Nh6 10.Qh3 Nf5 11.g4 Nxd4 12.cxd4 Qb6 13.Bg2 Nc6 14.Qh6 Nxd4 15.O-O Bd7 16.Qg7

In all five examples, it is black to move from the final position. Also, in all five examples, castling long/queen-side is perfectly legal and likely one of the best moves. BEFORE the game(s) is/are saved into the database, Fritz allows black to castle long (and it is at or near the top line in infinite analysis mode.) AFTER saving the game into a database, closing the game window, and re-opening the game from the database, Fritz treats castling long as an illegal move.