Author: Ralph E. Carter
Date: 16:11:27 10/23/98
Go up one level in this thread
On October 23, 1998 at 18:39:42, Ralph E. Carter wrote: >On October 22, 1998 at 23:16:23, Komputer Korner wrote: > >>On October 22, 1998 at 02:52:15, Ed Schröder wrote: >> >>>>Rebel 10 under Windows 95. >>>>PII 300, 28 Megs hash >>>>Game in 5 minutes. >>>>All defaults. No EOC. >>>>I am 1700 ICC blitz. >>> >>>>! I reached an easily drawn rook and pawn ending against Rebel 10 ! >>> >>>>Then, as I moved my king back and forth, it gradually ran out of time. >>>>Is this right? Surely it should be able to move almost instantly. >>> >>>It does move instantly when having less than 30 seconds on the clock. >>> >>>>I have lost the game score. It was 128 moves long. >>> >>>Congrats :-) >>> >>>Rebel on any (short) blitz level may lose on time. If you are able to keep a >>>game for more than 100 moves there should be a little (possible) reward. >>>This behavior (losing on clock) is very human alike. It's done deliberately. >>> >>>You will not have that pleasure on any tournament level :-) >>> >>>- Ed - >>> >>>>According to testing, it is performing as expected, otherwise. >> >>That doesn't make any sense Ed. I know that Hiarcs 6 does the same thing, but I >>think that you are joking on this one. Giving the user a bonus for lasting >>100 moves doesn't do anything for the user's pride or chess. The real reason is >>that doing the proper time allocation code for very fast games is more >>complicated and thus it is low on your priorities. It would be impossible for a >>human to manage his time as well as the computer anyway. It would only be >>possible if there were different time controls for each side, and why would a >>user take pride in this if he had a handicap in time to begin with?. Therefore a >>computer should never lose on time and I consider it a bug. If you want to have >>the computer to be able to lose on time, at least give the user the option of >>invoking this during settings. >>-- >>Komputer Korner > > >1. It is absurd to claim that THE Ed Schroder does not have his time allocation >code right after all of these years. If this were the case, he could rip-off >the Crafty code! > >2. The higher goal is to simulate the best human play. So, in blitz, to move at >the speed of a fast human is the IDEAL. > >3. For blitz training, time scrambles should be realistic! > >4. I didn't know HIARCS does this too. This is another plus for HIARCS. > >5. It is a common opinion that Rebel and HIARCS feature very human-like play. >This is one small part of that. Furthermore: I don't like blitz personally, but I admit that a good time scramble can be exciting. I love the Genius programs, but against Genius 2, there were no time scrambles: it was hopeless. After a while, I quit playing blitz against it. If I get a tiny advantage against Rebel 10 (in the game above I had a pawn plus at about move 50), I can play on, and hope to win! Just as I would hope to win against a real Grandmaster in such a position! This is realistic! I think that anyone who likes zero-increment games should consider this.
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.