Posted by Jonah Gardner on March 24, 2016
The Michigan State Spartans' defeat at the hands of noted powerhouse Middle Tennessee was a shocker for a number of reasons. MSU wasn't just the 8th 2-seed to lose to a 15-seed; they also had, according to our pre-tournament March Madness forecast, the best odds of winning the title. Even after the loss, Simple Rating System ranks the Spartans as the 3rd best team in the nation. And that's not to mention their star player, Denzel Valentine, who is the only person since 1994-95 to average 19 points, 7 rebounds, and 7 assists per game.
However, the most shocking thing about it, and a reason why so many brackets were busted on Friday, is that we simply aren't used to seeing Tom Izzo lose in March. Izzo has led Michigan State to 7 Final Fours since taking over in 1995, and many of those were with teams that appeared to have less talent and less regular season success than this one.
Thinking about Tom Izzo made me wonder if there was a way to quantify who the best March Madness coach in the NCAA was. After playing with some of the data, I came up with a very, very simple method for measuring how coaches have done in the tournament that I'm calling Coach Points.
Here's how it works: Coaches get 1 point for each appearance in the NCAA Tournament, 2 points for each Sweet Sixteen appearance, and 4 for each Final Four appearance. For every championship, coaches get 7 points plus the seed of their team. So winning the 2015 NCAA Championship was worth 8 points for Duke, since they were a 1 seed. On the other hand, a Syracuse win this year would be worth a whopping 17 points. In addition, I added 2 points for each win as a lower seed and subtracted 2 for each loss as a higher seed (this is since 1979 only, for coaches who have been around since before then).
To adjust for years as a coach, I also came up with Coach Points Percent. To calculate this, I multiplied the number of years a coach was in the NCAA by 15 (the total points a 1-seed would get for winning the tournament) and then divided their Coach Points by that number. CP% is like OPS or eFG%, in that, in theory, it can go well over 1. However, the scale made more sense to me since for every year a coach makes a deep run as a lower seed, they usually have many more where they didn't make the tournament or got knocked out early.
So, for example, let's look at how this would work with Bobby Knight. Knight made 28 tournaments (28 pts), 15 Sweet Sixteens (30 pts), and 5 Final Fours (20 pts). He won 3 NCAA Championships, one as a 3-seed (10 pts), one as 1-seed (8 pts) and one before 1979, but with a team that went undefeated, so it's safe to call them a 1-seed (8 pts). Since 1979, Knight won 6 tournament games as a lower seed (+12) and lost 15 as a higher seed (-30). The result is 86 career Coach Points and a .1365 CP%
There are, of course, numerous flaws with a method this simple. It puts far too much trust in the Selection Committee to accurately determine team strength, as opposed to using SRS or Net Rating. An ideal method would also find some way to adjust on-court performance for the strength of the players so that a team loaded with lottery picks and future stars, like the 2012 Kentucky Wildcats, isn't worth the same as an average 1-seed (538 has an approach that takes some of these factors into account).
This is just a starting point, but I do think the results are interesting. I crunched the numbers for every coach in the NCAA right now, as well as a few notable ones who recently jumped to the NBA. Here are the Top 11, in order of CP%. Note that, while I only say the school they're with now (or most recently), the results include their entire coaching careers:
Like a couple other long-tenured coaches, Boeheim's CP% gets knocked down a little because of how long he's been doing this. However, making the tournament 32 times in 40 tries is pretty impressive. Boeheim also has 10 wins as a a lower seed, 2nd most of any coach we looked at. On the downside, he coached the 1st 2-seed to ever lose to a 15-seed.
Because CP% is a rate stat, a newer coach can crack this list with one Cinderella run that goes deep in the tournament, which is exactly what Smart did. He's made the tourney 6 out of the 7 years he's been a head coach, but only advanced to the Sweet Sixteen once. And his 5 Ws as a lower seed are impressive, but he already has 3 losses as a higher seed, more than longer-tenured coaches like John Beilein or Gregg Marshall, who missed the list.
This model doesn't have a way to account for ethical concerns, so, on the numbers, Pitino makes the list at number nine. It's also worth noting that Pitino has more career Coach Points than Boeheim, despite coaching 10 fewer years in the NCAA.
The first of a couple of coaches who have moved onto the NBA who I added to our pool for curiosity's sake, Donovan only made the tournament in two-thirds of his season which is at the low end for coaches who made this list. However, he made up for it with a 9.5% championship rate, including one title that came as a 3-seed with the 2006 Florida Gators.
Cal's somewhat tortured March Madness history didn't come back to hurt him too much by this metric. He has 8 losses as a lower seed, but the 2014 Kentucky Wildcats' run to the Final as an 8-seed helped boost him to 6 wins as a lower seed.
Roy Williams is consistent. His teams made the tournament 92.9% of the time (only Mark Few's 100% rate beat that mark), made the Final Four 25% of the time, and won the title 7% of the time. His 11 losses as a higher seed are at the higher end of the list (other coaches with 11 include Bill Self and John Thompson), but it's not terrible considering how many highly-seeded teams he had at Kansas and UNC. And, if UNC wins it all, he'll pass Izzo in career Coach Points.
Here's something amazing: Coach K has been in the NCAA for 41 years and he has won the national title 12.2% of the time. That is how you end up getting the call to coach Team USA in the Olympics. His teams made the tournament 78% of the time and the Final Four 23.3% of the time. What holds him back in this metric is that he only has 5 wins as a lower seed and 20 losses as a higher seed. But that seems like nit-picking an all-time legend to me.
Brown, and two of the men ahead of him, show the Coach Point benefits of a shorter NCAA career with a couple really good runs. The embattled SMU head coach won 1 title in 11 years over the course of his NCAA career. That's a slightly lower rate than Coach K, but it came as coach of the 6-seeded 1988 Kansas Jayhawks. He also won 7 games as a lower seed, trailing just Boeheim and Izzo.
Stevens' 33% Final Four rate was the highest of any coach I looked at. He also never lost as a lower seed, while winning 5 times as a higher seed. Unlike Calipari and Pitino, the early results on Stevens' time in the NBA have been quite promising. Between him and Brown, it's possible that coaches with higher rates of tournament success actually do go on to have better NBA careers.
Izzo just barely edged out Stevens, thanks to a couple of factors. His Spartans made the tournament 19 times in 21 years, a rate that only Williams and Few beat. He also has 13 wins as a lower seed, by far the most of any coach I looked at. The only reason it's really close is because Izzo "only" has one title and it came with the top-seeded 2000 Spartans. How can anyone possibly top that kind of resume? Well...
Laugh all you want, but Kevin Ollie led the Huskies to a stunning championship, has gone 5-1 as a lower seed, and accumulated more Coach Points in 4 years than Mark Few has in 17 (he had 31). If Kevin Ollie coached the Huskies for 6 more years and never made another tournament, he'd still place 6th on this list. So, while the flukiness of the tournament will probably chip away at his numbers over time, his career is certainly off to a bright start.