Author: Jason Williamson
Date: 17:14:43 10/09/00
Go up one level in this thread
On October 09, 2000 at 05:04:42, pavel wrote:
>On October 09, 2000 at 04:16:18, Mike S. wrote:
>
>>On October 08, 2000 at 18:52:24, Vincent Diepeveen wrote:
>>
>>>On October 08, 2000 at 09:28:44, Mike S. wrote:
>>>>Should programs be able, or be allowed, to move in *zero* seconds of thinking
>>>>time?
>>>
>>>how about posting this in the mailbox of ICS admins instead of
>>>asking whether we want our programs to move fast, logically
>>>we want our programs to move as fast as possible. if timestamp only
>>>counts in 1/10 of a second, logically a program can move faster as that,
>>>so then it moves in 0 seconds. Logical. End of discussion.
>>>
>>>(...)
>>>Anyway i know your next post is about why programs are allowed to search
>>>so deep within a second (like several plies) where you make pure blunders,
>>>so that randomness should be added.
>>>
>>>Is it?
>>
>>Not at all. I wasn't talking about ICS play mainly (in fact, I hardly know what
>>timestamp is). What I have in mind, is the normal user situation ("offline"),
>>where he plays blitz games using a mouse, or even a sensor board. In the seldom
>>cases when he reaches a "dead draw" position at the end of a game, he probably
>>cannot agree a draw - because the program refuses to accept, and he must loose
>>on time because the program can play 10 or more moves in 1 or 2 seconds.
>>
>>Of course, anybody can choose a time control with increment. But I think it is
>>desireable, to have fair conditions under time controls without increment too.
>>
>>As I mentioned in another posting, I'm not playing blitz myself, and I have
>>always tried to express my opinion that playing computer opponents is generally
>>*not* unfair in any way. But I thought this zero, or nearly zero, thinking time
>>is an exception, chess programmers might want to think about and improve. I may
>>be wrong.
>>
>>This is not meant to keep a program optimally competitive for fast bitz games on
>>ICS, but more to create user friendly conditions for your (future) customers. I
>>could imagine to include such a minimum thinking time in a "etiquette" setting,
>>which would be applied if the opponent is human, and could be switched off is
>>desired.
>>
>>Regards,
>>M.Scheidl
>
>perhaps a better suggestion is to make the program understand drawn position
>well and accepts draw. So that in case of dead draw postions instead of
>contiuing the game, either the programs accecpts your draw offer or offers you a
>draw, I know few programs who understands drawn position relatively better than
>the rest; gandalf and crafty are the names i can remember right now.
>
>Pavel
And then maybe we can make humans that understand the same thing and don't try
to flag you in drawn positions.
JW
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.