Hi, I want to use decent CBR routines in a game I wrote. I know the theory about CBR''s well and also do know about some implementations in other fields... But I haven''t found a single hint that a single game used CBR-routines although there are some papers out there about the advantages this method would offer for games... So I would like to know if anybody here already used CBRs for AI-programming or if anybody knows of projects using them ... already the smallest hint will be helpfull ...

Well sorry to be not precise but I thought that CBR is well known ...
The AAAI already proposed them to be used in game for there learning possibilities and cause they offer so much more than e.g. A* or FSM ...
so you can use A* if the situation is clear and you can calculate easyly the effacts of all your actions (TicTacToe) ... CBR take unclear (e.g. fuzzy) values instead to choose from a list of known cases which is closest to the needed case ... then the algorithm will try to use the solution for the given case by adapting the old solution to the new case. If this atempt was succesfull the new solution is generally saved in the case databaseas a new sample of acting ...
So the advantage of CBR''s in games is clear: You need a startup database for many situations but while using this database the theory (and practical implementations like google) show that the system will realy learn/improove ...