Off the Charts
Written by: noreply@blogger.com (Rob Lowe (aka Henry Sugar))
Marquette's blowout victory against Cincy (http://www.crackedsidewalks.com/2009/01/this-just-in-mcneal-pretty-good-mu.html) was a great win. It was also notable for a few reasons.
Let's take a look at the four factors.
(http://2.bp.blogspot.com/_Xpuy4viivhI/SWLOJEIxT9I/AAAAAAAAAh0/k-GFFOQDdSI/s400/Cincy.png)
(http://2.bp.blogspot.com/_Xpuy4viivhI/SWLOJEIxT9I/AAAAAAAAAh0/k-GFFOQDdSI/s1600-h/Cincy.png)
141.2 was the second-highest (http://kenpom.com/expsked.php?&c=OE&team=Marquette) offensive efficiency for Marquette in 2.5 years, behind only the Coppin State game from last year (http://kenpom.com/expsked.php?y=2008&c=OE&team=Marquette). As has been recapped in many places, the key to this game was both the offensive performance (shooting 15-25 from three) as well as the defensive performance.
Although it was the offense that got most of the pub, I actually think that it was the defense that was the most important. Marquette's season average defensively is 93.6, and this game was well below that. In fact, the eFG% was the fourth lowest (http://kenpom.com/expsked.php?&c=O_eFG&team=Marquette) in 2.5 years, which by itself would have been enough to win the game.
(http://4.bp.blogspot.com/_Xpuy4viivhI/SWLQMfRmGlI/AAAAAAAAAiM/SiVoD7emdzg/s200/defense+vs+cincy.png)
(http://4.bp.blogspot.com/_Xpuy4viivhI/SWLQMfRmGlI/AAAAAAAAAiM/SiVoD7emdzg/s1600-h/defense+vs+cincy.png)Still, it wasn't just the defense alone that keyed the victory. Obviously, Marquette was dynamite behind the arc. This was also an off the charts performance, with the highest offensive eFG% in the last 2.5 years. I consider this especially impressive because Cincinnati came into the game as the #10 country at limiting eFG%.
(http://3.bp.blogspot.com/_Xpuy4viivhI/SWLRWp0MprI/AAAAAAAAAic/rqejzFbsSxs/s320/offense+vs+cincy.png)
(http://3.bp.blogspot.com/_Xpuy4viivhI/SWLRWp0MprI/AAAAAAAAAic/rqejzFbsSxs/s1600-h/offense+vs+cincy.png)Combine offensive and defensive performances that are off the charts, and you end up with a soul-crushing victory against a Big East Opponent. The big question... is this just a statistical anomaly, or a sign of the team improving at both ends under a new coach? The right answer is probably a little bit of both.
Given the improvement and wins, how do Marquette's chances look for the rest of the season? I'm not expecting a lot more than my New Year's Resolution (http://www.crackedsidewalks.com/2008/12/new-years-resolution-expect-less-from.html), but the news is better.
(http://4.bp.blogspot.com/_Xpuy4viivhI/SWLTutP9-vI/AAAAAAAAAik/rWOsv_EirdM/s320/predictions+-+1.5.08.png)
(http://4.bp.blogspot.com/_Xpuy4viivhI/SWLTutP9-vI/AAAAAAAAAik/rWOsv_EirdM/s1600-h/predictions+-+1.5.08.png)Pomeroy presently has us finishing with a most likely record of 11-7 in conference (http://kenpom.com/sked.php?y=2009&team=Marquette&t=p). Right now, I see a most likely view of 10-8 in conference (21-10 overall). If you're an optimist, there is an >80% chance of finishing with a record of 9-9 or better, with a 66% of finishing with ten wins or more. There's even a 13% chance of getting to 13 wins based on the statistics today. If you're a pessimist, there's a 45% chance that the team ends up at 9 wins or less.
Win on Wednesday, and the team practically ensures itself a record of 0.500 or better, with a solid track of finishing with a better record than last year.
http://www.crackedsidewalks.com/2009/01/off-charts.html
I know that the KenPom formulas can't account for teams that just "give up" after realizing that the difference in points has become "insurmountable." But I'm sure it's embedded somewhere in the defensive stats.
Yeah, there's a way to do it by un-correlating the offense and defense, but that's usually done on a macro (full season) level.
Also, it goes both ways. Marquette gave up a lot of offensive rebounds, but many of those were at the end of the game. Cincy made a number of shots during garbage time too.
If anything, I'd almost argue the eFG% on defense would be even better. Hard to say without writing 5000 lines of code to analyze PBP data (like Hoya Prospectus).