I got to thinking today: Who does Kentucky most resemble over the last several years in terms of Ken Pomeroy's offensive and defensive adjusted efficiency numbers? What I mean by that is, based on what we see right now, and assuming that the season ended today, what team that went deep into the NCAA tournament wound up with similar numbers over the season.
As you would expect, there is a pretty good correlation between Pomeroy's Pythagorean Winning Percentage calculation and the eventual champion, (70%) but it is far from perfect. But it does give us some pretty good data on which teams are actually likely to compete for the title.
Instead of using that calculation, I wanted to look at the adjusted offensive and defensive efficiencies, and the difference between them. That difference is more or less a competition-adjusted average margin of victory. That margin can be based on one of three situations — a strong offense, a strong defense, or a good balance of both.
Of course, there is the mandatory disclaimer of sample size, which is very small — only ten years. I'm not trying to create a rigorous mathematical model or proof or anything, just try to glean some interesting trends from the data we have. It's probably worth exactly what it costs.
So where is Kentucky right now? As of today, they look like this:
- Offensive efficiency: 118.1 (10th)
- Defensive efficiency: 95.9 (39th)
- Delta: 22.2
- Orientation: Offensive
It's no secret that Kentucky is offensively gifted, but defensively challenged. That's been a season-long thing, and even though they made defensive hay against Texas A&M, let's assume, just for the sake of conservative thinking, that they aren't going to improve much with better competition in that category.
So here's a list of the OE, DE and spread of the last 10 NCAA champions using Kenpom.com's figures:
|2009||1||North Carolina||122.4||92.9||29.5||O||North Carolina|
What you see here are the offensive efficiency, defensive efficiency, and the difference between them. Beside that, I have the best team in each year in terms of OE, DE, and delta if there was a team other than the champion. I also designated each team as O (offensively stronger), D (defensively stronger), or B (balanced) based on the difference between their ordinal rankings in each category. If the ordinal rankings are different by ten or more (and that difference is totally arbitrary), then the higher category is considered "stronger" for this exercise.
For reference purposes, the average numbers for a generic national champ based on these numbers look like this:
- OE: 119.14
- DE: 90.04
- Delta: 29.1
- Orientation: Balanced
As you can see, most years, the champ is usually the team with the best combination of efficiencies. This makes sense — Pomeroy's predictive model wouldn't be worth much if this weren't the case most of the time. But we do see cases where teams that looked "statistically inferior" won the tournament. You remember 2011, when Kentucky knocked off tournament top seed Ohio St., giving 3-seed Connecticut a shot. Illinois made it to the final game in 2005, but lost to UNC, the #2 statistical team. In 2004, the #1 and #2 teams were on the same side of the bracket and Connecticut knocked Duke off in the National Semifinal.
What is also interesting is my rather arbitrary orientation of the team. Balanced teams have won seven of the ten NCAA Tournaments. Offensive teams have won three. Defensive teams have won none.
So how does Kentucky compare? Well, in 10 years, only one team has won the NCAA tournament with a OE-DE in the same ballpark as as Kentucky's is right now — 2011 UConn. Kentucky's current performance level roughly equates to a #3 or #4 seed in the NCAA Tournament (although teams with better numbers have been seeded lower, like Pittsburgh last season, who was a #8). Although we have seen teams, namely UConn and Florida, win from #3 seeds, it's only 30% of the time that a winner comes from other than a #1 seed in this sample, and 20% for seeds other than a #2 or #1 seed. It's also interesting that only twice — Florida in 2006 and Pittsburgh in 2003 (who was out of my 10-year range) — has the top Kenpom performer not been a #1 seed.
What does this mean? Well, at this point, not a great deal. It does give you an idea, directionally, where Kentucky is right now in relation to other teams who have won the NCAA Tournament. It's a fair bet that Kentucky, by the end of the year, will be better than they are right now, but will that be good enough? Nobody knows, and these numbers won't tell us.
But could Kentucky win it all if these numbers don't change? Sure, it's possible. We have a small sample here, and I'm sure if Pomeroy were able to extend his numbers across the life of the modern tournament, you very likely would find an example where a team with even worse numbers won.
The best possible outcome, of course, is for Kentucky to get on a tear, start defending better, and then matters will take care of themselves. The Wildcats are in range of becoming an NCAA championship team despite the fact that they have not quite lived up to our elevated expectations of them to this point. I think that's a very good thing.