Survive and Advance

By Matthew Newton

It is known that March Madness, and the subsequent bracket picking that entails, is the epitome of annual unpredictability in sports for an entire month.  It remains clear that even the most consistent and strong teams from the regular season can disintegrate and implode in an early exit.  Upsets in the first round are inevitable, but that fact is not the focus of this article.  If your team manages to escape the first round, the margin by which that victory was achieved may be more important than you think.  Merely surviving and advancing might not be enough in the long run.  Getting pushed to the limit in the first round by a low-seeded opponent could be suggestive that, contrary to what you may have been thinking all season long, this is not your year. 

I ventured over to allbrackets.com, where the results of the last eight decades of NCAA basketball tournaments are recorded and archived.  The goal was to collect various data points regarding how highly-seeded teams tend to fare in the overall tournament after winning a relatively close game in the first round against a significantly lower-seeded team.  To find the most relevant sample to the current era of college basketball, I decided to use the past ten seasons (before the 2017-18 season) to collect my data (2008-2017).

First, I examined the first round margins of victory for the teams who reached the Elite Eight or further in the tournament.  I decided that the Elite Eight would be a reasonable base point for success because even a one-seed reaching the Elite Eight would likely not be a disappointing run.  In analyzing the 80 teams who have made the Elite Eight, as well as the teams who advanced past that round, there were some very telling results. The average first round win margin for Elite Eight teams over the past 10 years was 19.0 points.  Furthermore, the average margins of victory in the first round for Final Four teams and National Championship game teams were 21.9 and 21.2 points, respectively. Finally, the average first round win differential for the last 10 national champions is 27.7 points.  One might contest that these numbers are skewed, provided that it is essentially a given each year that most one seeds will obliterate their 16 seed opponent by a very large number, thus raising the average. Tossing aside the averages then, one crucial statistic is revealed: in the past 10 years, only one team (2014 Connecticut) has won the national championship after winning their first round matchup by less than 15 points (they beat 10 seed St. Joseph’s by eight points in overtime).  Considering this data, a conclusion is beginning to form that there is more at stake than elevating blood pressure for the fans when the low-seeded mid-major is taking the high-seeded blue-blood to the wire in the first round.

The second statistic I investigated focused on one and two seeds who lost in the second round (Round of 32), and if their first round game played just two days prior was perhaps a foreshadowing of their disappointing defeat. When scrutinizing the losses of one and two seeds in the first weekend of March Madness, some dominant programs make frequent appearances: Villanova, Duke, and Kansas to name a few.  It makes sense given the amount of times that these programs produce teams that achieve a one or two seed in the tournament.

Putting the phrase “the bigger they are, the harder they fall” in perspective, it is a truly shocking spectacle when these basketball powerhouses are defeated this early on in the tournament.  From 2008 to 2017, there have been 76 (out of a possible 80) one and two seeds that have made it to the Round of 32. Out of those 76 teams, 17 teams, or about 22%, failed to reach the Sweet 16. From those 17 teams, I sought to see if their Round of 32 defeat was indicated by a poor performance against a weak team two days earlier.  What I found was somewhat mixed results. The average margin of victory for one and two seeds in the Round of 64 over this sample was slightly more than 21.5 points with a fairly wide standard deviation of about 10.5 points. Only four out of these 17 teams failing to make the Sweet 16 fell more than one standard deviation below the Round of 64 victory margin mean.  Although the three teams with the lowest Round of 64 margins all lost in the next round, it does not appear that a poor Round of 64 performance necessarily indicates a tournament favorite will lose in the next round.

Another data set I explored made a convincing argument.  I reviewed every one through four seed from the last 10 tournaments, and singled out any team that won their first round game by 10 points or less.  I then analyzed the result of those teams’ tournament runs. This analysis provided the most significant conclusions. Over the past 10 seasons, a total of 35 teams out of a possible 160 seeded one through four won their first round games by 10 points or fewer.  Of those 35, just two teams reached the Final Four, which is 5.7%, a stunning rate when considering these teams are some of the highest seeded teams in the field. In fact, 70% of the Final Four teams in the past ten years have been seeded one through four, which means 17.5% of all one through four seeds make the Final Four (28 out of 160).  Neither of those two teams won or even reached the Championship game.

Additionally, I wanted to see if margin of victory in the Round of 64 could be used as a strong indicator of games won in the rest of the tournament.  To do this, I made the same decision to limit my sample to one through four seeds (thus, leaving some teams that made deep tournament runs out), as teams seeded towards the middle are obviously more likely to not only win less games, but also win their early round games by fewer points.  I then created a scatter plot with the number of wins on the x-axis and average margin of victory in the Round of 64 on the y-axis (one through four seeds that lost in the Round of 64 were not included). The graph below shows the results:

Based on this graph, there is a clear positive trend for average margin of victory in round one and total games won in the tournament, with the outlier, albeit from a small sample size, being from teams that reached the championship game. This model says that, over the average of all the one through four seeds, an extra 2.5 points (starting at 11.5 points) in first-round victory margin indicated that a team won an entire extra game in the tournament, and it explains 57.7% of the variance.  This model is surely not predictive, and using first-round victory margin as causality for future success would likely be a futile exercise. However, it certainly makes me think that a strong first-round showing is a welcome sign for any team that has title aspirations, and may be something to evaluate if you’re going to make any future bets or brackets after the Round of 64 concludes.

While blowing a team out in the first round is not necessarily the cause for a team’s future tournament performance, a strong first round showing for any team, particularly those playing weaker opponents, is at least a good sign for a team that could be poised to make a run.  It isn’t surprising that teams that play well at the start of the tournament are more likely to play well throughout it. So when you’re watching March Madness next year, make sure to keep an eye on how all of the favorites perform on the first Thursday and Friday, and maybe update your predictions based on who powers through and who only manages to survive and advance.