Author: Christophe Theron
Date: 23:24:33 07/30/00
Go up one level in this thread
On July 30, 2000 at 19:11:24, Jason Williamson wrote:
>On July 30, 2000 at 16:43:57, Christophe Theron wrote:
>
>>On July 30, 2000 at 04:34:34, blass uri wrote:
>>
>>>On July 30, 2000 at 03:13:42, James Robertson wrote:
>>>
>>>>On July 29, 2000 at 22:11:11, Derrick Wilson wrote:
>>>>
>>>>This is not a new strategy, but a time-honored method to win against computers
>>>>in fast sudden death blitz. It is not cheating or taking advantage of the
>>>>computer as, if it wanted to, a computer can move faster than any human.
>>>
>>>I agree that it is not a new strategy.
>>>
>>>The poster did not use the words taking advantage of the computer but taking
>>>advantage of the poor program.
>>>
>>>The poor program cannot move faster than any human.
>>>
>>>Uri
>>
>>
>>Right. I think in this kind of closed position it is possible for the human
>>player to move a piece back and forth while the opponent is thinking, making the
>>actual thinking time of the human player 0.00s or 0.01s or 0.05s at each move.
>>
>>Even if the opponent (the computer) does the same, what happens eventually?
>>
>>The winner of the game is decided at random, after one thousands moves, when the
>>0.05s have accumulated.
>>
>>It's a general flaw in the rules of chess servers, and not just a problem for
>>the chess programs.
>>
>>
>> Christophe
>
>I don't know how many times I have seen Crafty or Little Goliath get this
>situation on the ICC, and most of the time the program would be using a lot of
>time, get down to about 20-30 seconds left after shuffling around doing nothing
>like the human, then suddenly it would play a comital move and break open a line
>or something and overwhelm the human. Sure lots of time the human gets a draw,
>or even wins but in actual fact I suspect it works less then 30-40% of the
>time.
>
>JW
Chess Tiger 13 will have several ways of avoiding the problem. The problem has
been adressed by changes in the evaluation function and in the time management
algorithms.
Christophe
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.