[eDebate] .01 or 100 Points, Please... (second attempt)
Richard A. Garner
Tue Feb 17 14:32:29 CST 2009
Shorter Version: New point scales are good, and should be universally
adopted. Additionally, the scales should probably be varied every six or
Longer, epistolary version:
To Anyone Who Runs A Tournament:
1. New Point Scales: Not A Fail--Now that I've tried a hundred (96) and a
hundredth (29.8) a couple of times, the .5 scale system seems ... awkward,
ludicrous, unwieldy, played out, no longer as good, wrong, misapplied,
annoying, cumbersome, even anachronistic. I've talked to some people (though
not enough to extrapolate too much), and almost all of them (all of them?)
like the variations. I think that it's time to move on from the old 30/.5
scale on a fairly permanent basis. Our judgments, right now, are more
subtle, and need a point scale that can express that nuance.
1a. Preemptive Thought: Anyone that disagrees with me agrees with JVAR, or
even coin-flipping. (Actually, I think complex statistical tools might be
awesome. See comment 1b.i. below.)
1b. Additionally, my real proposal: we change speaker point scales every
five or six or seven years. It doesn't matter, really, it just needs to
change every time people get really set within a certain scale. The problem
is a subjective one, based on a complex interactional system that varies
according to micro-feedback over, in debate terms, macro-scales (i.e.,
longer than the career of a debater, even many judges, e.g., four years).
Every time it calcifies around a numerical norm it should be
1b.i. I propose, then, a variation between different base scales over those
period as well (30, 100, 50. 37,777, etc), because psychologically, people
don't think well outside of base ten or base sixty. Since the problem is
psychological--and it's only a problem until we solve it--then our solution
should be based on the reason it occurs.
(1b.ii. In any case, hasn't science advanced to the point that we can use
statistical methods to solve these problems? And by science, I mean, the
infinitesimal calculus developed in the 17th century... Anyone done a
comparative data study on the statistical significance of speaker points vs.
JVAR, etc? Does Nate Silver read this listserve?)
Richard A. Garner
p.s. According to the tubez, not FAIL = FTW. This is, at least, what the
kids these days tell me.
p.p.s. I prefer hundredths of 30, but think it will only last four years.
Then we can switch to hundreds.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Mailman