March Madness is so exciting, in part because every game can be a team’s last. Every single game is Game 7. That being said, there is not much that decides winners and losers in the NCAA Tournament. One shot, one tip, one rebound, and as we saw this year, seemingly one millimeter of the rim. Because most of these teams are so evenly matched, the team that advances is many times heavily dependent on the luck of the basketball gods shining upon them. So, does luck really play a factor in the teams that make a deep run and win the national championship? We turn to the great Ken Pomeroy for assistance with his analysis of college basketball. He defines “Luck” as the difference in a team’s actual win percentage and their expected win percentage. What do we mean by expected win percentage? We can predict how many games a team should win based on how many points they score and how many they allow per game.
Like I said earlier, games are regularly determined by slim margins that sometimes end up with unfavored teams winning and favored teams losing. Looking at Pomeroy’s luck statistic, we see that teams with a luck rating above zero are winning more games than expected, and are thus considered lucky, while teams with luck below zero are losing more than expected, and are thus unlucky.
This year, Pomeroy’s ten luckiest teams are
|UNC Greensboro, .167|
|Georgia St , .143|
|Kent St., .129|
|Abilene Christian, .123|
|UC Irvine, .110|
|Nebraska Omaha, .109|
|Bold denotes NCAA bid, Italics denote NIT bid|
Three of these teams made the NCAA Tournament by winning their conference championship. And two also won the conference’s regular season title, while the other finished second. Seven of Georgia St.’s 13 wins in the Sun Belt came with a two-possession victory or smaller. That luck led them to the 1-seed in the tournament. The Anteaters of UC Irvine won three OT games before upsetting Kansas St. by two possessions.
Even though these teams were fortunate and can be said they were somewhat “good at winning close games”, there was no guarantee that they would continue to play at the level that represents their record.
Conversely, let’s take a look at the ten (plus one) unluckiest teams of 2019.
|Boise St., -.192|
|Little Rock, -.149|
|Colorado St., -.111|
|Western Illinois, -.106|
|Penn. St., -.102|
As expected, there are no NCAA or NIT teams on this list. Northwestern (#73 in Kenpom overall) and Penn State (#40) certainly had the talent to gather more wins, but lost close game after close game in the Big Ten. Penn State lost nine key games this year decided by four points or less (or OT). While I am not convinced that a team has a statistical “clutch factor”, the chances of losing NINE games like that are not too favorable. Reverse a couple passes, shots, fouls, and rebounds and we could be looking at an NCAA bid here.
UC Irvine, Georgia St., and Abilene Christian were good but certainly lucky, and Penn State and Northwestern were great on paper but unlucky in their bottom line. A team with a decent record and good luck rating is a coach clinging to his job. A team with a great record and a good luck rating is a team you don’t want to face in a tournament. A team with a great record and a low luck rating is one you really won’t want to face, if their luck turns around.
Now, how about the luck ratings of teams who have made some noise in the dance? Let’s start with THE Cinderella. Who had the second-best luck rating of 2018? That’s right, it’s UMBC, the only 16 to ever beat a 1. Don’t say that didn’t have anything to do with the upset of Virginia. Who was number 22 that year? No, it’s not a trick question, it’s Loyola of Chicago. Middle Tennessee in 2016? Number 21. 2010’s #9 Northern Iowa over #1 Kansas? Number 85. These teams were average to good, but found the luck when it was needed the most, like UMBC cashing in from 3 against the Hoos.
When we take a look at the last ten national champions, we find that seven of them had positive luck ratings, meaning that the winners of the NCAA Tournament won more games than they should have based on their points scored and allowed, and were thus incredibly lucky. In a single elimination tournament like March Madness, luck plays a significant role in determining a matchup between close teams. The luckiest champion, as one would most likely guess, was the 2014 #7 seed Connecticut Huskies, the 23rd luckiest team that season with a rating of +.08. With two-possession or closer games over Saint Joseph’s, Iowa St., Michigan St., and Kentucky in the national championship, the Huskies pulled out a bunch of close games that could have gone either way. Other lucky champions include the Anthony Davis-led Kentucky team in 2012, the Duke 2015 champions with Jahlil Okafor, and of course the Kemba Walker UConn run in 2011. However, we can also identify the more “worthy” national champions who were unlucky throughout the season and still won the title. Louisville’s 2013 banner was vacated, but 2009 North Carolina and 2018 Villanova were two of the more dominant teams of recent memory, with each team winning every NCAA tourney game by double digits, which is incredibly impressive. These two teams certainly didn’t need luck to win it all!
Is winning these close games a skill? If it was, Coach K’s Duke squads should be ranked lucky year in and year out, but in reality, they have been not. Coach Krzyzewski was 135th this year and 271th a year ago. Calipari’s Kentucky squad was 98th this year but was just 199th last year. Even Bob McKillop’s “no one and done” team was 55th this year, 318th in last year’s NCAA appearance, and 259th two years ago. So, it seems the “luck” stat is just that, luck.
So, is it better to be lucky than good in the NCAA Tournament? In a single elimination “best-of-1” tourney, unless you can seriously separate yourself from the competition, better teams on paper will fall while other champions will rise.
This year, ALL FOUR of the final four teams have a positive luck rating. The final #1 seed remaining, Virginia, leads the group. A top-100 luck team has won the tournament five times last decade! The average luck rating of a winner the last ten years is .0273, close to where Auburn stands. This makes it clear that it doesn’t take just skill to win the NCAA Tournament. Virginia’s luck continues, and Duke’s just ran out.