A few people told me I should be posting on this topic. I plan on going into it in greater scope, in private, with R4's coordinators. Also, I should point out that I'm not trying to disparage any of our opponents. This is the tournament format that CBI RCT for region 4 at Bucknell used: - 12 teams were on the schedule. - Four rooms were on the schedule. - These teams were divided into two brackets (Red and Blue) of six by random draw. - Red and Blue played a round-robin within themselves, with two teams in each bracket playing each other in a sequestered round. Red played in rooms 1 & 2 and Blue played in rooms 3 & 4. - The records of the twelve teams (ties broken first by average differential) were arranged top to bottom. The six best records went into a Gold bracket; the six worst records went into a Silver bracket. - The Gold and Silver brackets played round-robins within themselves, with two teams in each playing a sequestered round on different questions than the non-sequestered teams used. Gold played in rooms 3 & 4 and Silver played in rooms 1 & 2. - The top three teams in the Gold bracket and the top team in the Silver bracket (ties broken first by differential) advanced to single-elimination playoffs. The schedule, as outlined above, presents the following issues: - Random draw for large fields works more reasonably, but with a field of such a small size, random draw increases dramatically the chance that the two brackets will be of unequal strength -- that is to say, the average team in one bracket will be stronger than the average team in the other, reducing the reliability of comparisons between the brackets. This did not appreciably happen at 2001 RCT, but it could very easily have. - With four rooms running, the RCT coordinators were willing to run sequestered rounds and two round-robins. A full round-robin among all 12 teams would have guaranteed all the teams an extra game to play, would have been more easily adjustible in the event of a no-show, would have eliminated the need for the calculations or meeting for division into Gold and Silver, and would have taken the same net time. - Arranging the use of rooms such that Red and Blue each had their own rooms, and then Gold and Silver each had their own rooms, leads to a situation in which one-quarter of the field only ever sees two moderators, and another quarter of the field only ever sees the two other moderators. Given that different moderators tend to read at different average speeds, and that individual statistics are taken in total points, this calls into question the entire validity of individual performance statistics as presented, and to a lesser extent the value of point differentials. This can be corrected for by dividing the individual scores and point differentials not by the number of games played, but by the total number of tossups heard by each individual team. [more]
This archive was generated by hypermail 2.4.0: Sat 12 Feb 2022 12:30:43 AM EST EST