Author: Robert Hyatt
Date: 13:41:48 11/17/04
Go up one level in this thread
On November 17, 2004 at 16:12:43, Dann Corbit wrote: >On November 17, 2004 at 15:59:24, Robert Hyatt wrote: > >>On November 17, 2004 at 14:43:43, Albert Silver wrote: >> >>>On November 17, 2004 at 13:28:33, Robert Hyatt wrote: >>> >>>>On November 17, 2004 at 12:45:58, Anthony Cozzie wrote: >>>> >>>>>Well, it is rapidly becoming that CCT time of year. >>>>> >>>>>I just wanted to get a little bit of discussion going on who is going to attend, >>>>>format, potential organizers, etc. I emailed Volker Richey, who did an >>>>>excellent job the last two years, but I haven't heard back from him. It is >>>>>possible I used the wrong email (cct@vrichey.de), so if anyone knows a better >>>>>one please speak up. >>>>> >>>>>Some thoughts: >>>>> >>>>>1. Timecontrol. Last we had some problems with some _long_ games (Movei the >>>>>Ironman comes to mind), and the first night ended up being pretty late for the >>>>>guys in Europe. Perhaps it is better to to just have a Game/60 timecontrol? I >>>>>like to have increments myself, but at the same time it is nice to know that all >>>>>games _will_ be finished by time X. I'm sort of on the fence here. >>>>> >>>>>2. Tiebreaks. Some would call this a bit of sour grapes, but I don't like the >>>>>idea of a slow tournament being decided by a few blitz games. I would much >>>>>prefer a buchholz or SB ranking, or even just having a shared first place. This >>>>>isn't a world championship, so there is no reason we can't make as many people >>>>>happy as possible :) >>>> >>>>I've been involved in the tie-breaks multiple times now. I agree. Get rid of >>>>the blitz play-off. Let's just have co-champions. Without proper seeding, even >>>>the traditional tie-breaks are not so good. Or, have a playoff at the _same_ >>>>time control as the regular games, but only if there are two tied and not more. >>>>Then we could play two games to eliminate white-bias. >>> >>>I have a suggestion regarding this, which is exactly how I think the genuine >>>human WC should be decided (as opposed to this grossly unfair draw favoring the >>>champion). In the event the 2 games do not provide a winner, use the now defunct >>>Golden Goal system from World Cup (soccer) events: the first to draw blood wins. >>> >>>There may be an element of luck involved in who would get white or black in the >>>first game, but all in all I think it still allows a satisfactory champion at >>>the end. >>> >>> Albert >>> >> >>Sounds ok to me. Flip for color, and start playing until a game is won? >> >>Only downside is that it is certainly possible (and even probable) to have an >>extended tie-breaker. :) > >If the programs automatically rematch each other, you could just go to sleep and >see who won in the morning, if it goes on for too long. > >It's not inconcievable that it could take 24 hours to decide. > >It might also be interesting to take the top two programs and make the final >championship as big of a deal as the CCT itself. So that (for instance) you may >decide to do a 10 game match or something of just the two top programs. > >It really depends on how interested we are in really finding out which program >is strongest. If we just want to proclaim a champion, then you could even flip >a coin to see who wins. It's really not much different from playing ten games, >since it takes at least 30 to really get a good statistical feel for which >program is strongest, if they are fairly close in strength. If it wouldn't take so long, it would be ideal to take the top N, then have another event where they play for the championship. IE with 64, take the top 4, then play at least a 6-round event where it is a full round-robin... But that's a lot of chess...
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.