Author: Dave Gomboc
Date: 21:42:22 04/30/98
Go up one level in this thread
On May 01, 1998 at 00:14:02, Robert Hyatt wrote: >On April 30, 1998 at 23:54:10, Dave Gomboc wrote: > >>On April 30, 1998 at 18:35:18, Robert Hyatt wrote: >> >>>I don't think any of this is a problem. The proper way to do this is >>>thru a "middleman" application that acts as a "server". Both programs >>>connect to that referee program and send stuff to it. It then sends the >>>stuff to the other program. That way programs don't talk to each other, >>>they talk through a third party, and discovering your opponent's ID >>>would >>>be harder. And the referee could be modified to further disguise things >>>if need-be... >> >>I think that last bit is weak. :) A player *should* be informed who its >>opponent is. If it knows how to use this information (e.g. can model >>the opponent's evaluation function to search a more accurate tree, can >>attempt to reach positions the opponent will mishandle) it could be a >>big plus for the program. >> >>Dave Gomboc > > >without it, massive abuse happens.. everyone books up against everyone >else, and only plays winning lines against known opponents.... That's okay, I do this in real life too. :) Seriously, home preparation is part of the game. A well-implemented book learning scheme should help out here. One issue that would need some attention is distributing the learned data to other copies of the same software that are being tested. Dave Gomboc
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.