Author: Robert Hyatt
Date: 21:54:48 10/05/99
Go up one level in this thread
On October 06, 1999 at 00:34:41, blass uri wrote: >On October 05, 1999 at 22:30:15, Robert Hyatt wrote: > >>On October 05, 1999 at 22:03:23, Christophe Theron wrote: >> >>>On October 05, 1999 at 16:17:39, Robert Hyatt wrote: >>> >>>>On October 05, 1999 at 13:58:53, Christophe Theron wrote: >>>> >>>(snip) >>>>>I suppose 100 blitz games would be enough to measure the "PB off" problem in >>>>>Crafty? Then if you want to fix Crafty in this regard, I'm sure you would get >>>>>plenty of volunteers to test the new version... >>>>> >>>>> Christophe >>>> >>>> >>>>I don't consider it as "needing fixing". Crafty won't run on a washing machine >>>>computer either. It could. But it wasn't a plan of mine to do so. Running in >>>>a crippled mode is something some like to do. But I don't plan on wasting time >>>>trying to tune everything so that it works well in a mode I won't _ever_ use in >>>>any serious games... >>> >>> >>>Running on "crippled" computers has helped me a lot to improve Chess Tiger this >>>summer. I have done a lot of testing on a 386dx20 computer in August and >>>discovered a selective algorithm I have just added in Chess Tiger 12.0. Ask Shep >>>about the speed improvement he has just noticed in the new version of Tiger. >>> >>>Don't worry, with your state of mind, I think you are very well protected >>>against this kind of discovery. >> >> >>I don't know why the discussion has to 'turn ugly'... but I'm up to the task. >>If it takes a 'crippled' machine for you to find new algorithms or find old >>bugs, fine. I have a better testing methodology. Because such testing is not >>exactly a modern approach to developing software. It is just as easy to cut the >>search time by 90% rather than finding a machine 1/10th the speed. Exactly the >>same result... >> >> >> >> >>> >>>As for "will it be useful in serious games", keep an eye on the results of Tiger >>>and you will know. >>> >>>Thanks for your time and patience, Doctor Hyatt. >>> >>> >>> >>> Christophe >> >> >>I'll bet you a buck that your testing with ponder=off won't help you one iota >>in "real" games... Because you won't be playing with ponder=off in "real" >>games, just like the rest of us.. > >I believe that testing with ponder=off is going to help christophe >in "real" games. > >You can learn much from the games about positional problems. >You can also see some problems in time using. _PLEASE_ _READ_ _WHAT_ _I_ _WROTE_ _CAREFULLY_. I did not say that testing with ponder=off was bad. I said that testing/tuning a program to work well in ponder=off mode will _not_ help you win games with ponder=on. There is a _BIG_ difference between what I said and what you seem to think I said. > >Every case when you play with ponder=off is similiar to playing with ponder=on >when the opponent is using 0 seconds for his move and it happans often in >comp-comp games when the opponent expect your move. all fine and good. and 100% irrelevant to what I wrote. You read my comments about Crafty. Do you want to say that testing with ponder=off helps _more_ than testing with ponder=on? Because I said that testing with ponder=off and tuning the time allocation code for that mode will _not_ help with ponder=on. And I stand by that... > >If your opponent is deep blue chip and it is using 0.1 seconds per move(in ICC) >then you have practically almost the same situation as playing with pb=off No because no one plays real games with that kind of a time handicap. DB doesn't move in .1 seconds. It takes as long to move as my program does. It just sees a lot more. > >It can be practical when deep blue chip is going to be commercial in ICC and >someone is going to test it in ICC and give it less time to give the opponent >chances. > >Uri Nonsense. Tune a program for that rare occurrence? Why? How many test games to get the time usage right. To do better in exactly how many real games on ICC? Total waste of time...
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.