16) How are players ranked? (College Bowl)
Description
This article is from the College Bowl FAQ, by George Atendido aten0001@tc.umn.edu with numerous contributions by others.
16) How are players ranked? (College Bowl)
(Pat Matthews)
Like any sport, academic competition has a few statistics which are useful
in measuring player performance. The most common stat is Points Per Game
(PPG). The formula for PPG is fairly simple:
(10 * # of tossups answered) - (5 * # of incorrect interrupts) PPG = -------------------------------------------------------------- # of games played
This formula can be applied to individuals or teams. PPG is often used to
determine tournament all-stars, but it is very vulnerable to the "shadow
effect," which will be explained further below.
PPG is sometimes rewritten as "adjusted average." The only difference
between PPG and adjusted average is that adjusted average uses 1 and -0.5
instead of 10 and -5 as the weighting factors for tossups and interrupts.
Other commonly used stats on both the team and individual level are Interrupt
Ratio or Interrupt Percentage. Interrupt Ratio is defined as the ratio of
incorrect interrupts to tossups answered correctly. Interrupt Percentage is
usually defined as interrupts divided by the quantity interrupts plus
tossups, but sometimes as interrupts divided by total chances (interrupts
plus tossups plus "zeroes", or incorrect non-interrupts).
A useful team statistic is Bonus Conversion. It is simply bonus points
earned divided by bonus points possible, and is the best measure available
for team play on boni.
Other, less widely-used stats also exist to evaluate performance. Most of
these alternative stats attempt to counter the "shadow effect." Simply
stated, the "shadow effect" is the effect on individual stats of having
very strong or very weak teammates. The idea is that having good
teammates decreases a player's opportunity to answer tossups, and that
having weak teammates increases said opportunities. Thus, Player A from
Team A could have a significantly lower PPG than Player B from Team B,
yet still be a much better player than B because A has better teammates.
One such measure designed to counteract the "shadow effect" is the Points
Created (PC) stat developed by Pat Matthews and Clay Davenport. PC
attempts to normalize player scores so that the effects of teammates on
scoring is minimized, and players are ranked according to how many
points they "produce" above or below the average player.
A full explanation of the PC method is available on RTF format from the
FTP site (pc_explained_rtf). In addition, a barebones, formulae-only text
version is avilable via FTP (pc_formulae_text). The reader is warned,
however, that only the rft version is "canonical", and the text version
may have an error or two in the formulae.
Continue to:
My Books
