Author: Stefan Meyer-Kahlen
Date: 11:30:13 04/17/04
Go up one level in this thread
On April 16, 2004 at 13:33:49, Russell Reagan wrote: >On April 15, 2004 at 15:31:22, Stefan Meyer-Kahlen wrote: > >>* ucinewgame > >Hello Stefan, > >Why did you guys choose a 'new game' command as opposed to a command like >'result 1-0'? I think a 'result' command would fit better into your stateless >approach, and allow for book learning better than a 'new game' command. I agree that "result" makes it easier for the engine to learn, but "ucinewgame" also makes this possible but doesn't wreck the stateless design of the protocol that much. As we seem to disagree here it might just be a matter of taste or attitude, but here is my point of view. A "result" is obviously from a game. I don't like that the engine actually knows about a game, for me it just gets positions to search. "ucinewgame" is of course the same if you interpret it as "ok, the old game is over and now starts a new game". I don't really like this. I would rather interpret it as "ok, all the next positions I am sending to you to search belong to something different than those I have sent to you before." So for me "ucinewgame" should actually have been called "ucisomethingnew", which would had been pretty stupid of course. So "ucinewgame" allows all the "my-engine-needs-to-learn-guys" to learn easier, so they are happy, and it allows me to interpret it my way, so I am happy. I hope there is some sense in the text I wrote above. >Here are some examples where I think a 'result' command would be more clean and >stateless than a 'new game' command. > >Let's say my engine just won a game. Does it get the ucinewgame command >immediately after the game, or does it only get the ucinewgame command >immediately before the next search? What if the game ends, then the user closes >the GUI? Will my engine ever receive a ucinewgame command? If it doesn't, I >can't do any book learning. No, you don't get anything. If you really want to implement a good and smart learning feature somewhere I would never count on somebody to tell me when and what I have to learn but draw my own conclusions. This is not very difficult for a chess engine to do and makes you independent on any external influences. Stefan
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.