Author: Robert Hyatt
Date: 06:03:21 05/24/01
Go up one level in this thread
On May 24, 2001 at 02:13:39, Bruce Moreland wrote: >On May 24, 2001 at 01:59:14, Will Singleton wrote: > >>On Volker's site (http://www.vrichey.de/cct3), I noticed two Polish programs, >>Armageddon Chess and Butcher. If these are known programs, then ignore the >>following, and accept my apology. >> >>If these are unknown programs, and have no track record of server or tournament >>play, should they be allowed to play in CCT3? >> >>Will > >Another issue involves how the manual programs will play. Having dealt with ICC >manual computer operators for several years, I'm concerned that they realize >that they shouldn't take an active role in the games. > >For example, I'd like to make sure that they don't do the following: > >1) Select book moves for the program. The program should select its own moves >once the game starts, meaning that the operator should not be sitting there with >ChessBase for the first 25 moves. The operator should not after 1. e4 e5 2. Nf3 >decide that he wants to see a Latvian, and type f5. If he sets it up that way >in advance, no problem, but this kind of reactive thing shouldn't be allowed >during the game. Or the operator plays _manually_ for the first N moves, when the program goes out of book, the operator disconnects and then comes back on with an automatic interface. This has happened in _both_ previous CCT events. > >2) Control time allocation. One way to play with a manual program is to let it >sit there thinking all the time. The operator makes moves when he thinks it is >appropriate. If it's an easy move, it comes fast. If the program looks like >it's going to play a bad move, or if the position is tactically complex, the >operator lets the program think longer in the hopes that it will find something >better. It is hard to run a manual program on ICC without doing something like >this, because you can't just select the tournament time control and run, because >of operator time and lag. But the operator should be as fair as possible. One >possibility is to set the program to run at fixed time per move, and shorten the >time per move as necessary as the game goes on. Another possibility is to tell >the program to run at a slightly faster time control than is going to happen in >the tournament, and hope that you'll accumulate time rather than running out. > >I'm sure there are other things that we'd all agree the manual operators should >not do. > >bruce
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.