[eDebate] A Possible Solution to the Speaker Point Dilemma
Thu Nov 29 07:20:24 CST 2001
I've advocated the use of judge variance
scores for speaker awards, but I have
one concern worth mentioning.
Patrick McKenzie writes:
>For example, consider a typical judge who judges
>4 rounds and 16 debaters in a tournament.
Is this really the "typical judge" anymore?
I see a lot of six (6) to eight (8) round
commitments and a lot of two (2) and three (3)
round commitments. It may well be that four
(4) rounds is the median commitment, but
I suspect any system of judge variance will
have to take into account the skews created
when a judge only sees four (4) or eight (8)
debaters during the tournament.
While the judges with larger commitments will
have a sample size which makes z-score
calculations meaningful, judges with smaller
commitments will risk skews. Example: Judge X
judges one prelim debate as a favor to the
tournament, but the debate is a horrible
mismatch and results in two (2) 29s and two
(2) 26.5s? Dropping high and low z-scores
might take care of this, but if low-commitment
judges are more prevalent, the risk of skews
Another query: why don't we make decisions
about which 5-3s or 4-2s break on the basis of
opponent wins instead of a deeply flawed speaker
point system. Is there any serious dispute that
a 4-2 accomplished against opponents who had
26 wins between them is "worth more" than a 4-2
accomplished against opponents who had 20 wins
between them? Perhaps this just begs the question
of whether the speaker point system "needs fixin'"
because opponent wins will vary with high-low
pairing which is . . . dependent upon speaker
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
More information about the Mailman