Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: CCT 3 Seedings

Author: Robert Hyatt

Date: 13:02:29 05/25/01

Go up one level in this thread


On May 25, 2001 at 09:58:04, Frank Phillips wrote:

>On May 24, 2001 at 20:08:41, Peter McKenzie wrote:
>
>>Do we have a rough seeding list?
>>
>>I'm not too worried about exact positions, but where possible we should use a
>>reasonable ordering.   Poor seedings tend to upset the first few rounds, for
>>example if Fritz played Shredder in the first round then some unfortunate first
>>round loser would probably get to play Fritz (or Shredder)...
>>
>
>If seeded I care neither whether I am top or bottom, but do not understand why
>one of the best losing to another top program upsets the first few rounds.
>Perhaps because I am not even sure what this means.

In a Swiss, tie-breaks are pretty common.  If you miss the seedings badly,
then the tiebreaks are bad.  IE the #1 seed will (if the seeds are 100%
accurate) have the highest tie-break since he will always be paired with the
strongest possible opposition (unless you have too many rounds as we do in
CCT [with 32 players, 5 rounds is optimal]).  If the best program is seeded
as the bottom player in the top half, he will have weaker opponents round by
round than the #1 seed, assuming both win until they meet.



>
>As far as I can see from the tennis analogy, seeding would seem to be designed
>to ensure that the best only play each other in final rounds for the benefit of
>the best, spectators and business interests.
>
>Why is seeding preferred to random initial pairings for this event. Is it some
>necessary feature of the Swiss system.



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.