March 19, 2012 - by Jack Moore
***NOTE*** This is a Round of 64 entry in our inaugural Stat Geek Idol contest. This article was conceived of and written by Jack Moore of Disciples of Uecker. The opinions or predictions expressed below do not represent the views of TeamRankings.com, and are solely those of the author.
We hear the story every year: the co-worker/significant other/random person who just filled in their bracket on a whim goes on to win the bracket pool. Meanwhile, the real fans, the students of the game filled out brackets sitting in the middle of the pack – or worse – despite hours poring over available game highlights or advanced statistics like those from TeamRankings.com.
This isn’t to say those fans who spend multiple hours on their bracket don’t necessarily know what’s going on. Advanced statistical rankings, whether they’re those here on TeamRankings or those provided by gurus like Ken Pomeroy or Jeff Sagarin, generally have a good idea of which teams are better and even how much better.
But, as is made clear every March, simply knowing which team is better isn’t good enough.
This fact leaves people looking for underlying factors which could power teams in March farther than similarly-talented counterparts. We hear this all the time: Team X doesn’t have Quality A necessary to make a deep run; Team Y’s ability to do Quality B sets them up well for March.
One potential “Quality B” is balance. Although there can be multiple definitions of balance in a basketball sense, let’s use the balance between offense and defense for the sake of this exercise. Do teams who perform equally well on both ends of the court fare better or worse than teams who particularly excel (in the optimist’s viewpoint, at least) on one end of the court?
In this case, the conventional wisdom holds: balance is a good thing, or perhaps more specifically, a lack of balance can kill. In a one game scenario, it usually isn’t enough to matter, but it would be incredibly difficult for a severely off-kilter team to make a six-game run at the tournament title.
First, let’s define balance: for this exercise I defined balance as the absolute difference in ranks between a team’s raw offensive efficiency and a team’s raw defensive efficiency. These differences rank from an even zero (such as 2007 Rice, which ranked 210th in each, and 2012 Kentucky, which ranks 4th in each) to 328 (2009 Hampton, which ranked 7th in defensive efficiency but 335th in offensive efficiency).
The average difference sits at 90, with a standard deviation of 70. Using one-half of a standard deviation as a basis for ranges, I constructed six categories of teams: Hyper Balanced, Super Balanced, Balanced, Unbalanced, Super Unbalanced, and Hyper Unbalanced. The Hyper teams are more than one standard deviation away from the mean, Super teams are between one half and one full standard deviation away, and regular teams are between zero and one half of a standard deviation away.
Since we want to know if the knowledge of balance supplements a simple knowledge of how good a team is, we’ll compare the results of the teams in these six groups to the Vegas lines for their games. The results show that balance doesn’t tell us all that much – until we get into the really crazy teams.
Unsurprisingly, there are far more games containing teams with balance than those without it – as a team’s rank in a category is capped at 1, the best teams will usually fall in one of the balanced categories, and unbalanced teams have to be bad at one of the two categories. Still, considering team quality is already accounted for in the Vegas spread, we would assume each category would come in at (or near) zero against the spread, and that is not the case – although the first five categories effectively come in within a point of the expected Vegas line, teams denoted as “Hyper Unbalanced” has severely struggled in 46 NCAA tournament games, falling nearly three points per game short of their projected margin of victory (or defeat — again, I prefer the optimist’s viewpoint).
Unfortunately, this isn’t nearly conclusive enough data to decide individual matchups. Typically even three points aren’t enough to change our definition of who the favorite is, and the 95% confidence intervals presented at the bottom of the table show that enough other variables are at hand that we cannot simply make picks based on balance.
Can there be use for this type of analysis, though? If so, it comes in what kind of teams we can project for deep runs. This year’s poster child for severe imbalance is Florida, which spent a decent amount of the year ranked in the top 10 in the AP Poll. Sitting at 19th in the TeamRankings Predictive Power Ratings, the Gators could justifiably be a darling for a run to the Elite Eight in some brackets.
Florida’s 165-spot difference between raw offensive efficiency (fifth, 115.7 points per 100 possessions) and raw defensive efficiency (170th, 100.7 points per 100 possessions), puts the Gators into the unfortunate Hyper Unbalanced classification of teams. Although I would hardly change my pick for Florida’s second round game against Virginia (another imbalanced team themselves) solely based on this revelation, chances are the imbalance will rear its ugly head at some point in the tournament. In the case of Florida, this comes when the team stops making three-pointers and the offense falls to a point where the defense cannot keep them in games.
It must be stressed that secondary qualities such as balance must always be just that when predicting games: secondary. A team’s pure quality as defined by its per-possession play over 30-some games tells us much more than anything else. Come tournament time, however, the games get so tight that victory in a bracket pool demands more. In this case, go with the conventional wisdom – severely unbalanced teams typically just don’t have what it takes to make deep runs.
Printed from TeamRankings.com - © 2005-2020 Team Rankings, LLC. All Rights Reserved.