Author: John Merlino
Date: 18:16:02 04/18/04
Go up one level in this thread
On April 18, 2004 at 20:58:50, Bob Durrett wrote: >On April 18, 2004 at 20:36:56, Steve Maughan wrote: > >>I'm looking to implement the new ELO rating support in UCI 2. In the past I >>haven't given much thought as to how one could make an engine predictably >>weaker. Does anyone have any suggestions? After the little thought I have >>given the subject here's what I've come up with. >> >>1. Search to a shallower depth / shorter time. This seems somewhat >>unsatisfactory as it will ignore the other setting. >> >>2. Add random elements to the evaluation term - the bigger the term the weaker >>the play. >> >>3. Round off at the root. So to weaken the engine slightly you would have the >>root round off to 0.1 pawns, a further weakening would be in units of 0.5 pawns >>etc. This means that if a value came back as +1.23 the engine would regard it >>as +1.20. I like this idea since it is relatively simple to implement :-). >>However I'm not sure how effective it would be at actually weakening the engine >>(it may speed the engine up since there may be more cuttoffs!?!). >> >>4. Within the tree you could ignore some moves e.g. all losing capture after x >>depth. This is also interesting. It may closely match how humans think and >>err. Of course it is more complex to implement. >> >>Any other ideas. This could be an interesting area for discussion as it has >>virtually no commercial value (IMO nobody is going to buy an engine that is >>weak). I'd be interested to know of other idea. >> >>Regards, >> >>Steve Maughan > >Awhile back, I had suggested making the engine play the second or third best >move except that it would play the best move if the second or third move was a >real loser. > >You could reduce the playing strength some by requiring second best [except as >noted] and more by making it play third best [except as noted.] Similarly for >fourth best, fifth best, etc. > >It would still not necessarily play human-like if the above method were used. > >The real challenge would be to make it play truly "human-like." > >Bob D. Indeed. A very difficult challenge, particularly when designing personalities. Some of the personalities in the 1000-1500 range in Chessmaster have some very odd settings, causing them to play generally "human-like" chess for most of the game, and then occasionally make a very bad play (usually a horrible sacrifice -- pawn for piece typically). But, of course, we had to give the personality SOME number, and if the math said 1400, then so be it. This is not the engine's fault -- it is merely using the settings that it was provided. But with SO MANY variables available in Johan's engine, it is pretty easy to make personalities that play very odd (but not necessarily BAD) chess. 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.