There are a couple more interesting cases from the mailbag this week. Let jump right in...
Memphis and Kansas have crazy low defensive efficiencies, especially in comparison to end-of-year DEff numbers from 2004-2007. How did the best numbers in late January of those years compare and how far did those teams come back to the mean? Thanks!
One of millions of things I'm considering to improve the adjustments I make to the raw offensive and defensive efficiencies is to start using data from past seasons to help assess what's going on in the current season. Right now, my system assumes that future events are just as certain after ten games worth of data as they are after 25, which is a bad assumption. Over the past four seasons, only one team has finished with an adjusted defensive efficiency of less than 84, and that was last season's Kansas team at 82.2. Through Sunday's games, there are five teams below 84; common sense tells us that they won't all finish there.
There have been five teams with an adjusted defensive efficiency under 80 through January 27 since 2004:
Team 1/27 Final
07 Duke 78.8 85.6
06 Texas 79.0 88.7
04 Cincinnati 78.8 90.2
04 Louisville 79.6 84.1
04 Georgia Tech 79.7 85.4
It's a small sample, but it does give you an idea of the uncertainty in the current values. Cincinnati fared the worst, but this also includes the game where they essentially quit in the second half against Illinois in the second round of NCAA Tournament, giving up 92 points in 63 possessions by the end. Both Kansas and Memphis have better defenses than any of the teams shown here, but that doesn't necessarily mean that either will finish as the best defense of the past five years, either. If any team is going to snag that mythical crown, it will be Kansas. Because Memphis' defense won't have to be that good in most of its remaining games, I'm fairly certain their defensive stats will slip relative to Kansas' numbers over the coming weeks.
We don't need formulas to know that Kansas and Memphis play terrific defense. This method is more revealing about teams for which there is uncertainty. I've been meaning to devote a whole column to Wisconsin for a while now. I've been unable to pull the trigger because the theme would be how the national media has been ignoring them, essentially on account of their style of play and the preconceived notions with which we entered the season, yet the data screams out that this is the favorite to win their conference. It's a familiar story and I'm unsure how to give a special twist. (I should point out that our own John Gasaway had the Badgers sharing the league title with Michigan State. However, the Big Ten media didn't have them in the conference's top three and the national media had them as just the 40th-highest vote getter in the AP preseason poll.)
For those who haven't noticed, Wisconsin currently has the third-best adjusted defensive efficiency at 82.3. Let's look back at the teams over the past four years that had a value between 81.3 and 83.3 on January 27.
1/27 Final Best case Worst case
Average 82.3 86.6 07 Kansas (82.2, 1st) 04 South Carolina (91.3, 21st)
There are 13 teams in this sample, so I'm not going to list them all, but the possible outcomes are not much different than the group of sub-80 defenses. The Badgers might have the best defense in the nation, and there's a small chance they're fooling the stat sheet and they're not really in the top ten; just two of the 13 didn't finish in the top ten.
Those who continue to believe that the two best teams in the Big Ten are Indiana and Michigan State are drifting from reality more with each passing game. It's unlikely that Saturday's road loss to Purdue is the beginning of a Badger collapse. Though a home loss to Indiana on Thursday would put the Badgers back on the national media's backburner for a while, even that wouldn't be the end of their realistic chances of a Big Ten title. This is a team that scores plenty of points to support one of the best defenses in the land. So don't look for them to fade any time soon.
I really enjoy your work. I think you have brought some much-needed rigor to the study of college basketball. I was wondering if you could comment on the performance of Arizona this season under Kevin O'Neill and how it compares to the performance of Arizona teams under Lute Olson the last three or four years. Does your data shed any light on this question?
Yes it does. The conclusion is this: if you make a high percentage of your shots and have trouble on the defensive end, it's foolish to crash the offensive boards. Let's compare a few stats from 2007 with 2008.
Adj. OE Adj. DE OR%
Year Rank Rank Rank
2007 7 72 43
2008 7 50 274
Kevin O'Neill has stopped sending four players to the offensive boards, instead relying on one or two guys for second-shot opportunities. This, in turn, has allowed the defense to prevent transition baskets, which has improved the overall defense slightly without much impact on the offense. I think the fact is, though, that the half-court defense probably hasn't improved all that much, and therefore the notion that O'Neill is some sort of defensive mastermind is misguided.
Almost all of the improvement can be explained by the fact that Arizona is willing to concede defensive rebounds much more so than in the past. Those 2007 figures, by the way, are fairly similar to what's on the record for 2004-2006. It should also be noted that the Wildcats' tempo has dropped by about four possessions per game this season. It's just conjecture on my part, but I'd bet most of the slowdown is the increased amount of time opponents are using when facing a more organized defense.
This also provides one of the first interesting storylines to watch for 2009. If Lute Olson does return, will he inherit the O'Neill philosophy which seems to be much different than his own? If everybody returns, Arizona would be a top-five team in the preseason. The subplots to such a scenario are fascinating, but the reality is that at least one of Chase Budinger and Jerryd Bayless is probably going pro. Nonetheless, it seems that O'Neill has tapped into an idea that Olson should adopt regardless of the personnel available in '09.
While we're on the topic, there are a couple other teams that could be well served by the O'Neill mentality this season. Both Florida and Xavier fit the profile of having high shooting and offensive rebounding percentages with an inconsistent defense. If the Gators and Muskies want to have aspirations beyond just contending for their own conference's supremacy, then sacrificing a few offensive rebounds might be a small price to pay towards a defense that can allow them to play further into March.
Ken Pomeroy is an author of Basketball Prospectus.
You can contact Ken by clicking here or click here to see Ken's other articles.