Author: John Merlino
Date: 12:53:36 09/26/02
Here are all of the current confirmed bugs in CM9000. Thankfully, it is a much shorter list than it typically is after the program has been released for one month: 1) Playing a Rated Game against a RANDOM opponent can incorrectly adjust your rating after the game is completed. 2) Playing a Rated Game and adjourning/resuming it can incorrectly adjust your rating after the game is completed. 3) The old (and very rude) "no matter who I choose in a Rated Game I end up playing the Chessmaster" bug is back, although it appears to be much rarer than it was in CM8000. If you are one of the very few who get this (and it only occurs on Windows XP), you should avoid playing Rated Games until the patch. 4) Blunder Alert can, on rare occasions, announce a blunder, and when you ask "Why?", the program says you should have made the same move that you DID make. Attempts will be made to greatly minimize this behavior, but it cannot be 100% eliminated. If you get this problem fairly often, it is advised that you increase the Blunder Alert Threshhold. 5) There is a bug in the code to find the proper move out of endgame databases. If there is a White pawn on a2 or a Black pawn on a7, or in very rare en-passant cases, then personalities that have access to the EGDBs will fail to find the move in the DBs and switch to the search. In most cases, this means that the personalities will not make the WRONG move, but rather have to find the RIGHT move through the search. This, thankfully, is NOT an error in the data, but an error in the code, and has already been fixed. Also, this error ONLY applies to personalities making moves in a game -- the data in the Coach window will still accurately show each legal move and the correct evaluation from the DBs. 6) There is a bug (or, rather, more of a design flaw) in the Solve for Mate feature. Even if a mate is found, it is not announced until the search finishes the current depth. In most cases, this difference is only a few seconds. However, depending on hardware and the position itself, this difference may be several minutes. All other problems that have been reported are either unconfirmed or were found to be user error/confusion. Now, to the bad news. Because the development team is currently swamped with creating the French version of CM9000, it is almost certain that this patch will not be released for at least one month, and more likely not until the middle of November. Many thanks for your patience, and to all of your help in finding these problems and reporting them.... jm
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.