"My rule was I wouldn't recruit a kid if he had grass in front of his house.
That's not my world. My world was a cracked sidewalk." —Al McGuire

Thursday, February 25, 2021

...Versus the Field?

A regular discussion in college basketball circles this year is whether people would take Gonzaga and Baylor or the Field if they had to pick a National Champion. The logic behind it is that Gonzaga and Baylor are so much better than the other 339 teams still playing this season that it is more likely one of those teams will win than the rest of the teams still competing for the title.

Would you take Jalen Suggs & Gonzaga vs the Field?

Photo by Douglas DeFelice | Getty Images

Anyone that watched Baylor last night (thanks to Big 12 Network, not a lot of people) may be questioning the wisdom of such a wager as they slipped by Big 12 bottom feeder Iowa State by a 77-72 margin, but I decided to look into the numbers behind the champions. The following table shows all of the National Champs of the kenpom era, including their Adjusted Offensive Efficiency Rank, their Adjusted Defensive Efficiency Rank, the Adjusted Efficiency Margin Score and Rank, their Seed, and their Adjusted Efficiency Margin compared to Pomeroy's #1 Ranked Adjusted Efficiency Margin team that year.

In layman's terms, how good their offense is, how good their defense is, their overall ranking compared to other teams that season, their seed, and the disparity between them and the respective number one that year, which effectively tells us how far away from the #1 team a bracket picker can realistically go to find a potential National Champion. Also bear in mind this data is all pre-tournament data. We aren't looking at how teams ranked after they won six games, we are looking at where they were before they went on that run. First, here's the data from the past 18 National Champions:

Year Team AdjOE Rank AdjDE Rank AdjEM AdjEM Rank Seed AdjEM vs #1
2002 Maryland 4 12 26.8534 4 1 -7.1686
2003 Syracuse 18 31 19.9564 20 3 -9.8108
2004 Connecticut 11 10 25.8489 5 2 -5.879
2005 North Carolina 3 6 32.0028 2 1 -1.3079
2006 Florida 13 18 23.9172 6 3 -4.6484
2007 Florida 2 14 28.8619 3 1 -2.8752
2008 Kansas 1 3 33.9647 1 1 0
2009 North Carolina 1 37 28.5574 3 1 -0.9203
2010 Duke 4 5 31.5489 2 1 -0.9632
2011 Connecticut 21 27 21.703 15 3 -10.529
2012 Kentucky 2 6 31.7277 1 1 0
2013 Louisville 17 1 31.1411 2 1 -0.842
2014 Connecticut 57 12 19.1162 25 7 -11.4515
2015 Duke 3 37 29.3074 6 1 -8.1259
2016 Villanova 15 7 26.7471 5 2 -3.0569
2017 North Carolina 4 25 28.0091 3 1 -5.0447
2018 Villanova 1 22 31.4091 2 1 -0.7436
2019 Virginia 2 5 35.6552 1 1 0

Average 9.94 15.44 28.13 5.89 1.78 -4.08

A few pieces of data jump out. Here are the big takeaways:

  1. Top-21 Offense: 94% of the champs were in the top-21 in offense, with only 2014 UConn as the outlier (get used to that). Which means we will be looking for a champion in the 94th percentile of defensive teams or better.
  2. Top-37 Defense: 100% of the champs were in the top-37 in defense. While the average defensive rank isn't as good as the average offensive rank, there are no outliers here. That means our 2021 National Champ should be in the 89th percentile of defensive teams or better.
  3. Top-6 Adjusted EM Rank: These numbers vary greatly by year because teams at the top and the grouping at the top is disparate, but 83% of the champs finished the regular season ranked in the Top-6 of kenpom. That's the 98th percentile of teams or better. The outliers were all still top-25, with 2003 Syracuse, 2011 UConn, and 2014 UConn breaking the mold.
  4. Top-3 Seed: 94% of the teams had a top-3 seed, with 2014 UConn again the outlier. Worth noting that 68% of the winners were 1-seeds, so you are in much better position picking that seed than all the others combined by more than 2-to-1 odds.
  5. Adjusted EM vs #1: 7 Champs were either the #1 team themselves or within 1 point of the top team, but that's just 36.8%, so not definitive. If we go to 5.02 (a number whose significance will be clear later) it rises to 57.9% of the Champs. If we go to 8.36 that's 78.9% of Champs, and to get to 100% we have to go to 11.45. 

So what does all this mean about who is a legitimate threat to win the National Championship? Let's take a look at the top-8 teams in kenpom currently. Before we exhaust that list, we will already be reaching teams that would be bigger statistical outliers than 2014 Connecticut:

Team AdjOE Rank AdjDE Rank AdjEM AdjEM Rank Seed AdjEM vs #1
Gonzaga 2 4 38.16 1 1 0.00
Baylor 4 9 33.13 2 1 -5.03
Michigan 6 12 30.73 3 1 -7.43
Iowa 1 75 29.79 4 2 -8.37
Houston 13 6 27.54 5 3 -10.62
Illinois 9 15 27.46 6 2 -10.70
Ohio State 3 86 27.29 7 1 -10.87
Villanova 5 64 25.00 8 2 -13.16

The bolded red scores indicate if these teams won a national title, they would be an unprecedented statistical outlier. Which means if you're looking at Iowa, Ohio State, or Villanova, the defense is simply too bad to expect a 6-game run and no team has ever won a title with a defense that bad. For Villanova, their Adjusted Efficiency Margin vs #1 Gonzaga is a bigger statistical outlier than 2014 UConn, which is a double-whammy. While people might point to teams like Houston and Illinois as viable picks, I would point to that Adjusted Efficiency Margin vs #1. While they would not be the biggest statistical outliers in the history of kenpom (2014 UConn strikes again) they would be more improbable than any team other than that one.

Ultimately, while it's not quite "Gonzaga and Baylor vs the Field" I do think it's safe to say "Gonzaga, Baylor, and Michigan vs the Field" is a safe bet where the smart money would be on one of those three teams over the other 338 options out there. They meet every criteria and while the Wolverines would be the fifth most improbable team based on that "vs #1" criteria, that's not nearly the crazy outlier that teams in the double digit disparities are.

Gonzaga really does deserve an extra mention. As noted above, the target number of -5.02 has produced 57.9% of the National Champions. So in terms of team success compared to the best team in the country, Gonzaga has better than 50/50 odds of winning the title. It's easy for casual fans to dismiss the WCC team, but their 7-0 Quadrant 1 record indicates they can play with anyone. Only Ohio State (8 Q1 wins) has more victories in that quadrant but the Buckeyes balance that with 4 Q1 losses. Gonzaga's SOS of 83 has been dragged down by the WCC, but it's still tougher than schedules faced by Baylor (203), Iowa (93), and Villanova (91), all teams on the top two seed lines that are never criticized for not scheduling tough enough.

That said, while Gonzaga's current 5.03 Adjusted Efficiency Margin lead on Baylor would break the record of #1 vs #2 by a wide margin, indicating Gonzaga is better against the field than any other team in the kenpom era, there is still risk. There have only been 4 teams to ever had an Adjusted EM margin of greater than 3 over the field. 2002 Duke (3.1778) lost in the Sweet 16, 2004 Duke (3.9222, the record) lost in the Final Four, 2008 Kansas (3.3894) won the title, and 2015 Kentucky (3.9082) lost in the Final Four. And that's why we love the NCAA Tournament, because even when you have prohibitive favorites, we are often surprised by the results.

A few notes on some of the teams not here. Pomeroy #9 Loyola-Chicago would have the worst offensive rank (#52) of any team to win the title. #10 Alabama would have the second worst offensive rank (#29) which feels crazy considering how much credit Nate Oats' offense has received. Could there be a team ranked #15 or lower? Anything is possible, but #15 Creighton has a -15.04 AdjEM disparity to Gonzaga while #25 Texas Tech has a -17.78 disparity, so seeing a team come from as far back in the kenpom rankings as those three outlier schools would be an insane level of statistical difference to overcome.

These numbers will fluctuate until Selection Sunday and this may be a topic we revisit (or an article worth reconsidering) once we have a bracket and it's time to make picks. But at this point, if you are picking a National Champion, you should really be looking at the top three teams as your only real options, and there's a pretty strong argument that any choice other than Gonzaga would be a somewhat improbable pick.

Here's the latest S-Curve:

1-Seeds: 1-GONZAGA 2-BAYLOR 3-MICHIGAN 4-Ohio State

2-Seeds: 8-VILLANOVA 7-ALABAMA 6-West Virginia 5-Illinois

3-Seeds: 9-Iowa 10-Houston 11-FLORIDA STATE 12-Oklahoma

4-Seeds: 16-Arkansas 15-Kansas 14-Texas 13-USC

5-Seeds: 17-Virginia 18-Texas Tech 19-Wisconsin 20-Purdue

6-Seeds: 24-Byu 23-Clemson 22-Creighton 21-Tennessee

7-Seeds: 25-Oklahoma State 26-Virginia Tech 27-Florida 28-Lsu

8-Seeds: 32-Rutgers 31-Missouri 30-LOYOLA CHICAGO 29-Colorado

9-Seeds: 33-Ucla 34-BOISE STATE 35-San Diego State 36-Maryland

10-Seeds: 40-Drake 39-Xavier 38-Minnesota 37-North Carolina

11-Seeds: 41-Louisville 42-Oregon 43-St. Bonaventure 44-Seton Hall 45-Colorado State 46-Stanford






Last Four Byes: Xavier, Drake, Louisville, Oregon

Last Four In: St. Bonaventure, Seton Hall, Colorado State, Stanford

First Four Out: Connecticut, Georgia Tech, Mississippi, Richmond

Next Four Out: Duke, Michigan State, Utah State, Memphis

Multi-bid Leagues

Big 10: 9

Big 12: 7

ACC: 6

SEC: 6

Pac-12: 5

Big East: 4

Mountain West: 3

AAC: 2

A-10: 2

MVC: 2

WCC: 2

No comments: