Offensive Variation and A Short Word about DefensePosted by nvr1983 on March 19th, 2009
When Ben Allaire isn’t drumming up meaningless college basketball statistics, he’s writing about the Virginia Cavaliers over at Dear Old UVa. RTC appreciates having Ben stop over this week to make some numerical sense of this year’s NCAA Tournament field.
What the analysis thus far has told you is how consistent a team is, without regard to how good they are. In theory, a team could be consistently abysmal and do well in its Pythagorean consistency. Granted, that team would be unlikely to make the big dance, but we wouldn’t have controlled for it.
Take a Pitt as an example. We’ve seen that Pittsburgh is a maddeningly inconsistent team in terms of offensive efficiency. As you probably have guessed, it’s a major function of their over-reliance on DeJuan Blair, but it’s a tad bit more subtle than that. Pitt is first overall in the country at offensive rebounding, again as a function of Blair’s obscene board work.
This reliance on offensive rebounding makes them susceptible to foes; if Blair gets in foul trouble, they have trouble producing points.
Furthermore, they don’t shoot that many free throws relative to their field goal attempt (in an attempt to get more offensive rebounds?). Free throws can be a consistent and efficient way to get the ball in the hoop.
Back to the main point: perhaps none of these reasons matter for Pitt to succeed. Maybe Pitt is so outstanding, even their bad times are good times.
That’s why I’ve “devised” a measure of offensive consistency relative to the offensive efficiency. Basically, I’ve just divided the standard deviation of offensive efficiency by its mean. Statistically, this is known as the coefficient of variation (CV).
The CV often used to assess how wide a distribution of numbers is. It has an added wrinkle here in that teams are trying to minimize the top (low variation) and maximize the bottom (high offensive efficiency). Thus, teams want this to be as low as possible.
Think about it like a dart board. The main goal is hit the bull’s-eye. Do you hit it once every ten shots, but otherwise are all over the place or are you consistently off by three inches? Really, you want to hit the bulls-eye often and when you miss, you want to be close to it.
Here are the top/bottom five teams seeded fifth or below in the coefficient of offensive variation:
|Rank||Top Performers||Off CV||Bottom Performers||Off CV|
|1||Oklahoma (B12) ||0.0954||Illinois (B10) ||0.1533|
|2||North Carolina (ACC) ||0.0980||Pittsburgh (BE) ||0.1501|
|3||Duke (ACC) ||0.1059||Memphis (CUSA) ||0.1438|
|4||Syracuse (BE) ||0.1093||Michigan St. (B10) ||0.1418|
|5||Washington (P10) ||0.1119||Louisville (BE) ||0.1417|
I guess this answers our conjecture about Pitt, but also raised a number of questions about Memphis and Louisville: two teams that many have winning it all. Will they have enough firepower to last eight games? You have to wonder.
I assume that no one is surprised to see Illinois tops at the bottom considering they’re (1) in the Big 10 and (2) that debacle against Pitt. Michigan St. has been pretty awful at times as well.
On the left side, we have OU. Jeff Capel has put together a great offensive group in Norman. Jeff, if you’re reading this, Charlottesville is lovely this time of year.
UNC is no surprise at #2. The secondary break is king, as is Ty Lawson (when he plays).
Duke is a surprising #3. I think most people think of Duke as been rather sporadic on offense because they rely on the three so much. Not so, they can hang with the top teams in the country on offense.
Lastly, I’d like to note that had we not included only the top five seeds, Siena would’ve been tops still. Fran McCaffery’s crew is really solid on offense. Fran, if you’re reading this, Charlottesville is lovely this time of year.
A final word on filling out your bracket, I’d hoped to do something with defense before today. Alas, it was not to be. I’ll still probably come up with something about it, but it won’t help you fill in your bracket. Be cautious with teams that force a lot of turnovers. From all the analysis that I’ve done, teams that force a lot of turnovers aren’t very efficient on D and are likely to get upset.