Final Four analysis: KenPom vs. seed success over the years
This is one of the more unpredictable Final Fours ever. But how does it stack up, and why are Final Fours tough to predict but not national champions?
If you go by combined seeding, this is the fourth most unpredictable Final Four we’ve had since the NCAA Tournament field expanded to 64 teams in 1985. And because of that, there’s even more reason to believe that No. 1 overall seed Florida will win the title, giving Billy Donovan three of them, essentially morphing him into an all-time college hoops legend by Monday night, should the best odds transpire. No matter your rooting interest, you’ve gotta admit: that’s freaking amazing.
But is Florida in an unusual position? How does this year’s Final Four crop stack up against the past 11 years? And why 11, you might be asking? The 2002-03 season was the first for KenPom.com, one of the most well-known advanced-statistics sites/metrics. It’s not the be-all, end-all, no way (Pomeroy would just as quickly remind you of this). But it’s proven to give us among the most accurate gauges of teams’ capabilities and greater standing among the house of cards that is the field of 68. For example, take No. 11 Tennessee. The Vols barely snuck into the field, having played in the First Four. They go on to make a Sweet 16 run (something that’s happened in three of four years since the First Four began) and almost beat No. 2 Michigan in the process. (It wasn’t a charge!)
Seeding can cloud presumption, but Tennessee’s three-game spree shouldn’t have been too much of a shock. The Vols were ranked 13th in KenPom on Selection Sunday. Yet in seeding, a massive disparity. We get this every year with a handful of teams. This happens in part because KenPom is predictive, while seeding is entirely reflective. It’s why I wanted to examine the past 11 Final Fours and see how seeding compared to Mr. Pomeroy’s inventory. The KP rankings in this piece are all taken as the teams stood in his echelon every Selection Sunday.
First, here’s what the average seed of the Final Four has been since 2003.
Now let’s see what KenPom’s average ranking of the Final Four teams has been. The reason for choosing the Selection Sunday timeframe is fairly obvious: NCAA Tournament success will only increase a team’s placement in the rankings. This is a truer indicator of a team’s predictable/unpredictable run to four (or in VCU’s case, five) more wins.
You look at the two charts, and it’s nice to see the patterns are fairly close. The rises and falls line up. In 2011, we had an absolute aberration — but 2008 also gave us something even more unlikely. The odds of getting all four No. 1 seeds back to the end of the road seem slimmer by the year. Cherish that ’08 foursome, because it might be two decades before we see it again.
The average KenPom ranking of every Final Four squad since ’03: 11.38. This year’s four-team average: 14. So, less-predictable than usual but less volatile than two of the past three seasons. Regarding seeding, the average seed of a Final Four team since 2003: 3.1. This year: 4.5. Only 2006 and 2011 were wilder.
For comparisons sake, and to see how seeding can contrast with how a team is ranked in Pomeroy’s formula, here are the differences over the years, including 2014. Tennessee this year is a team that stands out. George Mason was a controversial inclusion prior to making history in 2006 — yet the Patriots were much better that year than VCU was in 2011. Both 11 seeds, both made the Final Four, as you know. How about 2005 MSU?
National champs in capital letters.
What about the title winners? A few things on this. First off, only five times has the No. 1 team in Pomeroy’s pre-tourney rankings gone on to make the Final Four: 2004 Duke, 2005 Illinois, 2008 Kansas, 2012 Kentucky, 2013 Louisville. The latter three wound up winning the whole shebang. That alone can speak to the fickle nature of this amazing event. This year’s elite victim was Arizona, which came so close. Of the past 11 champions, here are the pre-tourney rankings. You’d have thought (at least I did) UConn in 2011 would’ve been the lowest-listed team, but in fact it was third-seeded Syracuse. Pomeroy: yet another Melo hater.
The average KenPom ranking of a national champion since 2003: 5.64. This year, Florida obviously winds up as the closest to that average.
And on the flip side, the lowest-ranked KenPom teams to make Final Four. The only one to make the title game was 2011 Butler, and it beat a team ranked even lower heading into the matchup:
79th: 2011 VCU
45th: 2011 Butler
35th: 2012 Wichita State
25th: 2014 UConn
22nd: 2006 George Mason
22nd: 2010 Michigan State
Patterns show the tournament is giving us good but not great finishes, merely from a seeding/ranking standpoint. This doesn’t preclude exciting endings, but there’s an interesting conflict at play. While the Final Four offers up enough unpredictable occupants that they go beyond outlier status, the teams actually winning the whole thing aren’t shockers at all.
The difference between winning four games and winning six is like a canyon’s jump. Look at who’s won since 2003: Syracuse, UConn twice, UNC twice, Florida twice, Kansas, Duke, Kentucky and Louisville. All of those programs are top-10. All of them. And three are back in the Final Four this time around. Many believe seeding doesn’t matter. When it comes to the final game, that’s an accurate assertion.
Motley, who averaged 17.3 points and 9.9 boards, is projected as a borderline first-rounde...
Sharpshooter's return figures to send the program to a place it has been only once before
Are you buying?
Gary Parrish and Matt Norlander also discuss UNLV's recruiting class
The five-star big man is part of a surprise late-period recruiting coup by Marvin Menzies
It's time for random observers to stop being outraged by players' decisions