Over the past couple weeks I've received several emails like this:
Could you please publish the efficiency margins of last year's Sweet 16 or Elite Eight? I find your method very interesting, but would like some comparative numbers with relation to last year's brackets.
For anyone just tuning in, Lee's referring to tempo-free scoring margin, the difference in a given team's points scored and allowed per possession.
Lee, this time of year the most important thing to realize about efficiency margin is that the selection committee has no idea what it is. A good efficiency margin can tell you a lot about a team, but if the committee puts that team on an imminent collision course with a one-seed, the pretty numbers won't matter. The seed that a team receives Sunday night is just as important as the EM they've built over the past two months.
Not that efficiency margin can't be used as a predictor, mind you. Here are last year's top eight major-conference teams in terms of EM:
Best Efficiency Margins, 2007
Conference games only: ACC, Big 12, Big East, Big Ten, Pac-10 & SEC
Pace: possessions per 40 minutes
PPP: points per possession
Opp. PPP: opponent points per possession
EM: efficiency margin (PPP - Opp. PPP)
Pace PPP PPP EM Tournament Result
1. Kansas 71.0 1.13 0.89 +0.24 Lost to UCLA in Elite 8
2. N. Carolina 74.4 1.13 0.96 +0.17 Lost to Georgetown in Elite 8
3. Georgetown 62.6 1.14 0.97 +0.17 Lost to Ohio St. in Final 4
4. Ohio St. 62.2 1.10 0.94 +0.16 Lost to Florida in National Final
5. Texas A&M 65.8 1.13 0.98 +0.15 Lost to Memphis in Sweet 16
6. UCLA 63.1 1.10 0.95 +0.15 Lost to Florida in Final 4
7. Wisconsin 61.3 1.07 0.93 +0.14 Lost to UNLV in Second Round
8. Florida 66.5 1.13 1.00 +0.13 Won National Championship
Not bad, right? Even a BCS-driven data set that doesn't include Memphis still pegs 75 percent of the Elite Eight. Wisconsin notwithstanding, the only teams these behemoths lost to were those Tigers and each other. Even a defending national champion that took February off (the Gators were actually outscored over their last five SEC games) before turning it on NBA-style in March couldn't hide from the trusty old EM. Yay, efficiency margin!
Well, not so fast. The first thing to note here is that last year was weird. With the exception of the Badgers' loss to UNLV, no one- or two-seed lost a game until the Elite Eight. You know that annoying person in your office pool who simply picks the higher seed in every game? Last year that person did very well.
The teams listed above all had beautiful efficiency margins, sure. Even more important, though, is the fact that, with the exception of third-seeded Texas A&M, they were all one- or two-seeds. So just remember one thing on Monday morning when you're penciling in some early hunches on your brackets: Looking at major-conference teams over the past two seasons, there's a stronger correlation between tournament wins and seeding than between tournament wins and conference winning percentage or efficiency margin.
Which is simply to underscore the fact that the committee has tremendous power when they assign seeds. The difference between being a three- or a four-seed, for instance, can be huge. It's also why plain old vanilla wins and losses really do matter.
The committee's seeds the past two years have tracked better with conference wins than they have with something weird and obscure like efficiency margin. So if you had a choice between being a three-seed with a very unimpressive EM (e.g., Oregon last year) or being an underappreciated five-seed with a very good EM, you would of course take the better seeding, the one that keeps you away from the one-seed for another game. Skeptics of mere stats, raise a glass!
Nevertheless, there might be a couple of insights that efficiency margin can offer. Consider these teams this year:
Pace PPP PPP EM
Texas 65.3 1.09 1.00 +0.09
Connecticut 67.7 1.11 1.02 +0.09
Michigan St. 63.1 1.06 0.97 +0.09
Mississippi St. 67.7 1.05 0.96 +0.09
West Virginia 64.5 1.07 0.99 +0.08
Marquette 69.1 1.06 0.98 +0.08
Stanford 63.2 1.05 0.97 +0.08
Washington St. 58.5 1.09 1.01 +0.08
Clemson 69.4 1.07 1.00 +0.07
Notre Dame 72.2 1.11 1.04 +0.07
Indiana 64.8 1.08 1.01 +0.07
Ohio St. 63.5 1.02 0.95 +0.07
USC 65.3 1.06 1.01 +0.05
Arkansas 68.8 1.04 0.99 +0.05
I'll be watching these teams very closely. To my eyes there are some very tough outs there. Here's the strange thing, though: over the past two tournaments, the performance of teams like these, with major-conference efficiency margins in the 0.05 to 0.09 range, has been atrocious. Such teams are just 17-19, which, interestingly, is only a hair better than the 17-21 record posted by teams in the next category down, with EM's between zero and 0.05. When I first saw these records, I thought it must be because the better EM teams were running into high seeds sooner than the lower EM teams.
In fact, the opposite is true. In 11 of 19 instances, major-conference teams in the 0.05 to 0.09 range were sent home by a lower-seeded team. The lower-quality teams in the zero to 0.04 range, by stark contrast, lost to lower-seeded opponents just five times in their 21 losses. That's striking; if it continues this year it'll officially be time to put on the guessing hat and figure out what the heck is going on here.
The other potential tournament lesson learned from efficiency margin is more obvious but bears mentioning. If a major-conference team is outscored over the course of their conference season, they're probably not going to do real well in the NCAA tournament. It's common sense, sure, but tell that to the committee--these teams keep getting bids. This year's pool of potential sub-zero entrants includes:
Pace PPP PPP EM
Virginia Tech 72.2 0.97 0.98 -0.01
Maryland 72.7 1.03 1.04 -0.01
Miami 69.6 1.06 1.07 -0.01
Ole Miss 69.7 1.06 1.08 -0.02
Arizona St. 62.8 1.00 1.02 -0.02
Oklahoma 63.6 1.01 1.03 -0.02
Florida St. 67.7 0.99 1.05 -0.06
In the past two tournaments, major-conference teams with negative efficiency margins are 1-4. Here's one instance, at least, where EM might be almost as mighty as seeding. Not even the implausibly cushy five-seed given to surprise Big East tournament champ Syracuse in 2006 could mask the fact that they'd been badly outscored during the regular season by their Big East opponents. (The Orange were sent home in the first round by 12-seed Texas A&M.) The best that can be said of such negative-EM teams is that one of them, Indiana in 2006, was able to beat San Diego State by four. That's it.
I'll leave you with this year's top eight major-conference efficiency margins:
Pace PPP PPP EM
1. Kansas 69.2 1.16 0.92 +0.24
2. Wisconsin 60.2 1.09 0.91 +0.18
3. UCLA 64.9 1.13 0.96 +0.17
4. Tennessee 71.5 1.12 0.97 +0.15
5. Louisville 67.0 1.06 0.91 +0.15
6. North Carolina 74.7 1.13 0.99 +0.14
7. Duke 75.3 1.11 0.99 +0.12
8. Georgetown 62.6 1.04 0.92 +0.12
Adjust for reality, of course. Kansas plays in the Big 12 North and gets eight games against Colorado, Iowa State, Missouri and Nebraska. Wisconsin was blessed with eight games against Northwestern, Penn State, Michigan, and Iowa, yet had no road game against the Big Ten's third-best team, Michigan State. North Carolina had a key injury.
Still, the numbers here should probably be treated with some respect. Recent history says that where seeding and EM align, the chances are good that a given team will be joining presumptive co-entrant Memphis in the Elite Eight.
John Gasaway is an author of Basketball Prospectus.
You can contact John by clicking here or click here to see John's other articles.