If you’re normal, you probably spent your weekend socializing or catching up on sleep. If you’re like me – a nerd – you spent your weekend scraping 10 years of historical data and statistics in college basketball ahead of March Madness 2022. The question I attempted to answer: Which regular season team stats tend to correlate with success – and failure – in the NCAA Tournament?

I looked at nine March Madness tournaments (forgetting that the 2020 tourney was canceled due to COVID – sue me), and plotted every team’s regular-season data against how many wins they had in the tournament. I left out the one-and-done mid-majors (since their losses wouldn’t count as underperformance), but included the teams that won games to see if there was any way to figure out which could go on a run in 2022.

A disclaimer: As anyone who has taken a statistics class knows, correlation does not mean causation. Every year, there were major outliers and teams that throw all of the findings out of wack. So, when filling out your bracket, don’t take these trends as gospel – think of it as just an extra thing to consider when you’re trying to determine which teams can make a run. Individual matchups will be far more important than historical trends – but if you’re stumped, hopefully this can straighten you out. Or, at least, this can give you a sense of which teams typically go far in the tournament as you pick your Final Four.

When plotting March Madness wins against all team statistics, the stat that had the highest correlation – visually, the one whose trendline had the steepest slope – was team field goal percentage. Intuitive? Sure. “Make more of your shots, win more games” isn’t exactly a bold statement. But when looking at the data, this stat was the one with the fewest outliers for teams that advanced deep into the tournament.

Of the last nine teams to win it all, only three of them (Louisville in 2013, UConn in 2014 and North Carolina in 2017) ranked outside the top 32 in team field goal percentage. In that same span, only seven teams that advanced to the Final Four (out of 36) ranked outside of the top 100. And no team since 2011 has won it all without making over 45 percent of their shots in the regular season.

For the highly-successful March Madness teams that didn’t make a crazy percentage of shots, there was another trend: they all were very good at limiting opponents’ shot-making, which is the other highly correlated statistic – defensive field goal percentage. Louisville and UConn had the lowest two offensive FG percentages of the last nine champions, but each of them ranked inside the top 24 on defense (coincidentally, they both allowed opponents to shoot 39.8 percent). As an inverse example, Missouri in 2012 led all of the NCAA in field goal percentage (50.4 percent), but was outside the top 200 in opponent% – and got bounced in the first round as a two-seed by Norfolk State (who shot 66.6 percent from the field in that game and scored 86 points).

So, to sum it up: if you hit a high percentage of your shots, and limit your opponents to a low percentage, you tend to do better in the tournament. Again, fairly obvious – and often correlates with high seeds – but still important to understand. This year, Gonzaga and Arizona both rank in the top 10 of both categories – and are a worthy top-seeded championship matchup for your brackets. Kentucky is inside the top 50 in both. Kansas, Duke, and Baylor all rank highly in offense but are a bit further down the list on defense. Villanova and Auburn are the inverses, higher on the list in defensive field goal percentage.

The top 10 March Madness teams in shooting percentage differential, between offense and defense:

Surprisingly, three-point percentage was one of the less correlated statistics for teams that went far in the NCAA tournament. The last three champions (Baylor, Virginia and Villanova) all ranked inside the top 12 and shot over 39.5 percent from deep. But Michigan State (in 2016) and Creighton (in 2014) both ranked number one in the NCAA in three-point percentage, and didn’t make it to the Sweet 16 as top-three seeds. MSU’s loss was one of the biggest upsets in the history of the tournament, in the first round to No. 15 Middle Tennessee. Also, two champions (Louisville in 2013 and North Carolina in 2017) ranked outside the top 150.

As explained above, recent history has been kinder to three-point shooting teams. But teams that shoot more threes don’t typically have more success – in fact, the number of threes attempted per game is negatively correlated with March Madness success, if only slightly. Four of the last nine champions were outside the top 200 in threes attempted per game.

Opponent three-point percentage, however, did show a slight correlation. Teams that made it to the Elite Eight and beyond averaged a full percentage point better at defending the three than the “shame teams”. By shame teams, I am referring to top-seeded teams upset in the first two rounds. Again, not a huge difference, but if you’re looking at three-pointers, defense may take a slight priority. Baylor is the best top seed at defending the three, while Kansas, Kentucky, and Gonzaga trail not far behind in the top 50.

Offensive rebounds per game and assists per game showed little correlation with success compared to the average team. That’s not to say teams that do these things well haven’t gone on runs in the past: North Carolina actually won the title in 2017 ranking in the top three in both categories. But UNC ranked in fifth that very next year in both, and lost in the Round of 32. Less correlation just means there are examples more evenly distributed on each side of the success/failure spectrum, which has been the case here.

Free throw percentage, too, showed so little correlation that the slope was basically a flat line. There are extreme cases on both ends, even for Cinderella teams. Two 15-seeds have advanced to the Sweet 16 in the past 10 years: Oral Roberts in 2021 and Florida Gulf Coast in 2013. Oral Roberts ranked first in the nation with an amazing 82.1 percent, while FGCU was 256th at just 66.9 percent. Like threes, free throws are a much smaller sample size in a 40-minute game, and can wildly swing a team’s favors with one good (or bad) evening.

This leads to my final point: March Madness is really freakin’ random. Every year, great teams with great underlying numbers suffer baffling losses to mediocre squads. That’s what happens in a single-elimination format: small sample sizes lead to extreme variance. The beauty of March Madness is that any given team can have a great day and beat someone way better than them. It’s what makes the tournament so exciting – but predicting it so hard.

Whether you’re using these trends or not when predicting your bracket, don’t be too hard on yourself when your picks go horribly wrong. Often, a good process can lead to bad outcomes – and the winners can be predicated on luck. That said – and this is perhaps the most important trend of all – seven of the last nine winners have been one-seeds. If that holds, the best conclusion of this study is perhaps that Gonzaga and Arizona are your best bets when picking an overall winner – and overall winners count more than anything in most bracket formats.