Author: John Merlino
Date: 10:49:51 05/23/01
Go up one level in this thread
On May 22, 2001 at 21:19:05, Dieter Buerssner wrote: >On May 22, 2001 at 14:52:28, John Merlino wrote: > >>The WB protocol seems pretty clear to me. Here's what it says: >> >>---------- >>hard -- Turn on pondering (thinking on the opponent's time, also known as >>"permanent brain"). xboard will not make any assumption about what your default >>is for pondering or whether "new" affects this setting. >> >>easy -- Turn off pondering. >>---------- >> >>Therefore, your statement to turn off pondering "you have to send hard followed >>by easy" doesn't seem to be true. According to the above, if pondering is on by >>default, you SHOULD only have to send "easy" to turn it off. Perhaps the >>standard implementation among engines assumes that what you have said is >>correct, but the protocol states otherwise. >> >>However, I have noticed that WinBoard does send "hard" automatically at the >>beginning of every game, and THEN sends "easy" if pondering has been turned off >>in the GUI. I don't know if it HAS to do it this way, though. >> >>Either way, if the patch needs to be tweaked further, I will make sure that, at >>the very least, "hard" is sent automatically. If possible, I will add something >>in the GUI to allow the user to select whether pondering is on or off for that >>personality. > >Sorry, I was obviously confused. Nevertheless, I still think, that the CM method >is inferior, and I think my arguments are still valid. (And I also still only >know CM from log files). I think with the CM method of leaving the ponder status >to the default, will yield in strange tournament results. You may also consider >this statement (which probably was the source of my confusion): > >hard >easy >Sent in sequence to turn off pondering if xboard is not sure whether it is on. >When xboard is sure, it will send "hard" or "easy" alone. xboard does this >because "easy" is a toggle in GNU Chess but "hard" is an absolute on. > >And also especially this: > >hard >Turn on pondering (thinking on the opponent's time, also known as "permanent >brain"). xboard will not make any assumption about what your default is for >pondering or whether "new" affects this setting. > >I think, the last really shows, that CM is not right to not send either hard or >easy. BTW. My engine really implements easy as described by you (as a switch, >and not as a toggle), but sending hard and easy cannot hurt at all, and looks >superior to me. > >Regards, >Dieter Well, you're correct about WinBoard, and I think your argument is valid. Therefore, I have created a full GUI implementation of "pondering", in which each XBoard personality has a field in which you can tell CM whether to send "hard" or "hard/easy" after each "new". This, of course, is not exactly reverse compatible. What is going to happen is that, with all previously imported engines, EASY will be the default (because that field in the personality structure was set to zero -- meaning pondering=off). This will all be explained in the ReadMe for the patch, so hopefully people will read it! The testers are going to kill me, though.... Thanks for getting me into trouble. ;-) jm
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.