March Madness: Using Trends to Narrow Down the Eventual Champion in 2023
This has become one of my favorite March Madness articles to write every year because it helps keep expectations in check.
While we all know that March's magic can lead to turmoil at the start of the men's NCAA tournament, we also know that it's quite shocking to see upsets in the Sweet 16 and beyond.
With that in mind, our realistic pool of championship contenders should stay pretty small. How small? And how do we whittle things down?
Well, with data, of course.
At numberFire, we have detailed game-by-game projections and bracket tools that help predict this year's iteration of the tournament, but we can also use historical benchmarks of past tournament winners to see which types of teams tend to win the whole thing and help you find bracket success.
Which trends matter, and which teams fit this year? Let's find out.
Seed and nERD Data
nERD is a big part of the March Madness season at numberFire. Simply put, it's a one-figure number that measures the expected point differential for a team against an average opponent on a neutral court.
We use that to power rank teams from 1 through 363 -- but also to power rate the teams and figure out how big a gap there is between one team and another.
The tournament seeding process sets out to do the former but not so much the latter.
Digging back to the 2000 season, we have had 22 tournament winners. Of those 22, 16 (72.7%) were 1 seeds. We also had a pair of 2 seeds, three 3 seeds, and a lone outlier, the 2014 Connecticut Huskies, who were a 7 seed.
Here's where we can start to blend seed with nERD.
Yes, UConn was a 7 seed in the tournament. However, their nERD of 14.06 suggested that they were roughly equivalent to a historical 4 seed (whose nERD average is 14.08).
If the teams in the tournament were ranked by nERD, then, effectively, each of the 22 winners since the turn of the century would have been a 4 seed -- at worst -- with a heavy majority of them being 1 seeds.
The average seed for an eventual champion is 1.64, and the average nERD is 19.4.
So, again, a 4 seed or better is really the way to go, which automatically cuts us down to 16 teams if we trust the process.
As far as nERD rankings go here's a snapshot of nERD distribution by range for champions.
by nERD Rank
|6th or Worse
Of the past 22 champions, 9 were ranked as numberFire's best team overall for the season, and 19 of them were fifth or better.
The three exceptions to the rule were the 2003 Syracuse Orange (9th in nERD), the 2011 UConn Huskies (13th) ,and the 2014 UConn Huskies (20th).
Realistically, we're looking at top-10 teams going on to win, but we can make the cutoff a top-20 nERD score.
Offense and Defense
At numberFire, we use percentile rankings for team offense, defense, and pace.
A fun note each year for this piece is that offense is what seems to win championships -- not defense.
Since 2000, 21 of the 22 championship teams had an offense in the 89th percentile or better; those darn 2014 Huskies were in the 81st percentile. This means a top-70 offense at worst should be our target -- but more realistically it's around a top-40 offense.
Defensively, four champs had a defensive rating in the 81st percentile or worse, including last year's Kansas Jayhawks (76th-percentile defense). The lowest we should be dipping is around the 70th percentile, a top-110 unit.
While the cliche that "defense wins championships" isn't entirely false, offensive prowess has been a stronger predictor of championships than defense has.
We'll simplify it and look for teams with an 80th-percentile offense and a 70th-percentile defense.
Teams That Fit
If I filter our database for teams who meet certain criteria, these teams remain.
2023 NCAA Tournament Teams With a Qualifying nERD, Offensive Rating, and Defensive Rating
These teams have a nERD rank of 20th or better, an offensive rating in the 80th percentile or better, and a defensive rating in the 70th percentile or better. Note that four teams don't fit the 4-seed-or-better criteria.
|Alabama Crimson Tide||18.31||1||0.90||0.96|
|Saint Mary's (CA) Gaels||14.38||5||0.90||0.99|
|San Diego State Aztecs||13.33||5||0.85||0.82|
|Florida Atlantic Owls||12.26||9||0.92||0.93|
Notable names missing (and the reason):
- Tennessee Volunteers: 75th-percentile offense
- Gonzaga Bulldogs: 56th-percentile defense
- Marquette Golden Eagles: 60th-percentile defense
- Baylor Bears: 55th-percentile defense
- Utah State Aggies: 42nd-percentile defense
- Xavier Musketeers: 34th-percentile defense
- Arkansas Razorbacks: 69th-percentile offense
- West Virginia Mountaineers: 67th-percentile defense
- Kansas State Wildcats: 21st in nERD and 75th-percentile offense
- Indiana Hoosiers: 28th in nERD
- Virginia Cavaliers: 34th in nERD
If you really like one of these teams, you're probably within the outlier parameters, though it's likely better to target a team that's not in the table due to a moderate defense than to trust a weaker offense.
As always, though, you're free to do whatever you wish with your bracket. After all, it's your bracket.
But historical trends alone say that one of these eight teams will be the 2023 NCAA champion:
- Houston Cougars (note: senior guard Marcus Sasser missed the team's AAC championship game with a groin injury)
- Alabama Crimson Tide
- UCLA Bruins (note: they lost junior guard Jaylen Clark for the season, and freshman center Adem Bona is dealing with an injury, as well)
- Connecticut Huskies
- Texas Longhorns (note: senior guard Timmy Allen missed the Big 12 tournament)
- Purdue Boilermakers
- Arizona Wildcats
- Kansas Jayhawks (note: coach Bill Self missed the Big 12 tournament due to health concerns)