Re: Athlon Sports on scoring efficiency

["Jim Meninno" <jam@london.com>]

----- Original Message ----- From: "Thomas Murphy" <tfmiii@worldnet.att.net>

> I guess you don't understand the concept of a normal distribution or a
bell
> curve. Statistically we know that if we have enough instances of a
player's
> performance those instances when plotted should make a curve shaped like a
> bell with most instances occurring near the average and fewer instances
away
> from the average eventually tapering off to nothing. So in your
hypothetical
> if a player averages just under the 'cut off' that means slightly over 50%
> of that players instances will lie below the cut-off and 40+something
> instances will lie above the cut-off. There is nothing 'arbitrary' about
it.
> If you can actually find an example like you describe you'll be able to
> re-write the texts on statistics!
>
> Best wishes, TomM

It's obvious that you know a lot more about statistics than I do, but I do
know what a bell curve is.  I apologize for the sarcastic tone of my first
reply, you've put a lot more thought into this than I first believed.

When I called the cut off point "arbitrary" I meant that there is no
specific reason for putting it where it is.  I don't have the original post
anymore, but my recollection of the rationalization is that any performance
that exceeds the league average efficiency helps your team win, and any that
falls short of it does not.  I think that's an entirely unsafe assumption.
There are countless factors other than the league average efficiency to
consider.  For instance, if Milwaukee allow their opponents to score much
more efficiently than New York, then their own players need to score more
efficiently to "help their team win".  Conversely, a player on the Celtics
may help his team win against New York with a much less efficient game than
in a game against Milwaukee.  Another example would be that it would be much
harder to score efficiently in a game where the referees were not calling
many fouls.  A below average efficiency in that game might actually do the
job.  Applying the same standard to every game on the calendar seems bound
to produce misleading results.  This is why I feel like this rating is
claiming to measure something it can't possibly measure, which is the number
of games that a player's scoring was good enough to help his team win the
game.

It also seems misleading to me that extraordinarily good games are included
in this measure, but poor ones are not.  We can't expect Walter McCarty to
have another 6-6 3pt. shooting night like he did a couple of years ago, but
his rating would be helped by that game.  Why is this method better than
taking the median?  That would seem to more fairly account for both positive
and negative outliers.

Jim