Author: Ed Schröder
Date: 01:06:16 08/11/04
Go up one level in this thread
On August 11, 2004 at 03:34:13, Ed Schröder wrote: >On August 10, 2004 at 23:08:35, Mark Young wrote: > >>On August 10, 2004 at 19:13:44, Ed Schröder wrote: >> >>>On August 10, 2004 at 12:43:19, Ed Schröder wrote: >>> >>>>On August 10, 2004 at 11:54:33, Mark Young wrote: >>>> >>>>>On August 10, 2004 at 11:15:30, Ed Schröder wrote: >>>>> >>>>>>On August 10, 2004 at 09:08:13, Mark Young wrote: >>>>>> >>>>>>>Ed, Pro Deo 1 is having a problem with understanding time controls. I have >>>>>>>posted Pro Deo 1.O games below in the current test I am running. As you may know >>>>>>>in my test at 40 moves in 2 hours and 20 in 1 hour time controls. Pro Deo 1.0 >>>>>>>only scored 1/2 point out of 5. >>>>>>> >>>>>>>When I went to look at the game I found out why.... From move 1 to 40 Pro Deo >>>>>>>1.0 would use the time control correctly. >>>>>> >>>>>>>Between moves 40 and 60 Pro Deo 1 would only use 1 sec per move. >>>>>> >>>>>>This is a correct observation already known from Rebel 12. You are probably >>>>>>playing 40/2hr 20/1hr 1hr/all. The engine can't handle this and starts to play >>>>>>at 1 second per move, no surprise it has only 0.5/5. >>>>>> >>>>>>I suggest you to use 40/2hr 20/1hr 20/1hr. >>>>>> >>>>>>Sorry for all the wasted time! >>>> >>>> >>>> >>>>>Ed, Then the problem is worse in Pro Deo. Look at my Time control from the games >>>>>below. I used 40/2hr 20/1hr 20/1hr. See Below. >>>>> >>>>>****[TimeControl "40/7200:20/3600:20/3600"]**** >>>> >>>>I see. >>> >>> >>>>I assume you are using the CB interface. If so zero all items of the 3th time >>>>control. I never had any problems. >>> >>>This is also not sufficient. >>> >>>I think it is better to wait for a bug-fix version, this is a bug with very >>>unpleasant consequences. >>> >>>Ed > >>Yes, I agree. No need to test if you are going to fix and/or update the program. > > >I did some testing and it seems that changing the following wb2uci.eng parameter >does the trick: > >LevelExtend = none -> LevelExtend = strict > >I will do some more testing with other time controls as well, in the meantime I >think it is safe to restart your testing with that change. > >My best, > >Ed Mark, I would like to take this back, just leave wb2uci.eng unchanged. Instead of that make time-contol-1 equal to time-control-2 and the problem disappears. So: 40/2hr+40/2hr or 20/1hr+20/1hr Zero the 3th time control. As far as I know problems like these don't happen in ChessPartner. I will make a note on my pages. My best, Ed
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.