Wednesday, March 17, 2010

What stats leading to winning and losing??

After looking at stats in the aggregate, I thought I would go back and look at stats for each game.  The goal is to figure out what stats were closely linked to winning and losing, and which ones were not.  Please note that somebody who actually knows statistics might correlate this stuff....with me, you get more of a shade-tree statistician look.

For reasons you will see in a minute, defensive performance had a bigger role in winning and losing then offense did.  What I did was take our average for each stat, and then figure out our record when we were better than our average and worse than our average.  The stats you see below is the different in record between better and worse than average, expressed in games.

By far, the single biggest and most influential stat is point per possession allowed.  BG was 12-1 when it allowed fewer PPP than its season average and 2-15 when it did not, a 12 game difference.  The next strongest test was True Shooting (which measures FG, 3FG and FT shooting) which had a 9-game difference.  That one is a little disingenuous since you can't defend against FT%.  Floor percentage (percentage of possessions scored on) was +7 as was effective FG%.  (When BG allowed eFG below 45%, they were undefeated).

Other tests revealed a less striking difference.  Note, in particular, that tempo did not appear to play a big role.


Switching to offense, the first thing to notice is that the strongest offensive measure only made an eight game difference.  On the whole, you can see in the chart below that there are much fewer long columns in the offensive chart than the defensive chart.

BG was 11-4 when it was above its average in offensive rebound % and only 3-12 below.  The other interesting one was steals allowed, which showed a seven game difference.  It is almost TOO interesting, in fact, and seems like something random might be going on there.  Floor percentage was also key for the offense.  Scoring was less important (+5) than for the defense.



Anyway, like I said, there are probably more sensitive ways to test this, but I'd say that a stronger defensive performance would have yielded us more wins than more scoring.  As I noted in the aggregate stats analysis, our defense was clearly below our expectations for the style of ball we want to play.

No comments :