Author: Brian Richardson
Date: 16:05:13 11/24/04
Go up one level in this thread
On November 24, 2004 at 18:03:59, Uri Blass wrote: >On November 24, 2004 at 15:31:12, Anthony Cozzie wrote: > >>On November 24, 2004 at 15:05:17, Brian Richardson wrote: >> >>>On November 24, 2004 at 12:15:44, Will Singleton wrote: >>> >>>>On November 24, 2004 at 09:37:21, Anthony Cozzie wrote: >>>> >>>>>On November 24, 2004 at 09:04:11, James Swafford wrote: >>>>> >>>>>>I've seen a couple of attempts to get some discussion started about >>>>>>CCT, but AFAIK nothing's on the books yet. >>>>>> >>>>>>Anthony- did you ever get up with Volker? He's fantastic at >>>>>>organizing these things; I hope he's willing to do it again. >>>>>> >>>>>>Was a date /format ever decided on? I'd like to play again this >>>>>>year, but that depends on the date (I can't play the weekend of >>>>>>the Jan 22nd/23rd). Of course, majority rules... >>>>>> >>>>>>I think a lot of folks will want to start lobbying for their >>>>>>favorite time control, but maybe we can start by agreeing on when >>>>>>we can play: >>>>>> >>>>>>1. If you are interested in playing, what weekend(s) would >>>>>> work for you (or wouldn't work for you) in the Jan/Feb >>>>>> time frame? >>>>>> >>>>>>2. What are your thoughts on stretching the event out over >>>>>> two weekends? Would you be less likely or unable to play? >>>>>> >>>>>>-- >>>>>>James >>>>> >>>>>I just got Volker's email from Ralf, and sent him an email not 5 minutes ago. I >>>>>agree he did an excellent job last year. Anything in January _or_ February is >>>>>fine with me; I have no social life worth talking about ;) >>>>> >>>>>I'd like to propse 60 2 as the timecontrol for CCT. The problem with a large >>>>>increment is that since we are playing with computers, there will always be *at >>>>>least one* game that goes out to 150 moves. So a 10 second increment will give >>>>>the average game 60*10 = 10 minutes, while delaying the tournament 150*10 = 25 >>>>>minutes. Therefore, I think keeping the increment small is a good idea, and if >>>>>you screw up your time allocation, that is your problem. At 60 2, we can pretty >>>>>much guarantee all games will be finished in 60 + 60 + 4*150/60 = 2:10, which is >>>>>about right IMO. >>>>> >>>>>I'd also like to point out (again) that I don't like tiebreak blitz games. :) >>>>> >>>>>anthony >>>> >>>>With a 60 2 time control, you might get more than tiebreak blitz games. :) >>>> >>>>45 10, with 9 rounds over one weekend, seems to have worked well for the last >>>>two events. CCT4 had 2 weekends at 60 10. 2 weekends makes it tough for some >>>>folks to participate, and I think we ought to try to continue to make it as >>>>accessible as we can. >>>> >>>>So I would vote for keeping the same format as last year. Late January or early >>>>Feb is fine. >>>> >>>>Will >>> >>>I also would prefer only one weekend. >>>Also, I don't think there is a significant difference in what moves most engines >>>would play between 45 and 60, but I do think it is very important to try to >>>finish each round in under 2 hours so the maximum number of rounds per day can >>>be scheduled. >>>Brian >> >>I am not so sure. 45 seconds per move isn't really very much; Zappa will >>probably only reach 12 ply in critical positions at that timecontrol. I still >>regularly see the best move change at higher ply. >> >>anthony I meant 45 minutes to start plus 10 seconds per move increment. These time controls have been used for several CCTs, IIRC. My thinking is that with most engines having an effective branching factor of about 3x (to search one ply deeper), that unless the time differences are also about 3x, it is better to save a little time in the interest of scheduling this event, which could have 50+ engines participating. > >1)By this logic no time control is enough. > >I am sure that you will still see the best move change at higher ply even if you >search 13 or 14 plies. > >2)If we assume that you have 1 hour per game and that the game takes 60 moves >then you have 1 minute per move and you practically can use more than it because >of pondering. > >Uri
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.