Predicting NCAA Tournament Results Months in Advance
Are preseason rankings accurate predictors of future success in the NCAA tournament?
Mirror, mirror on the wall, which team will win it all?
As March approaches each year, every college basketball radio station, TV broadcast, and news article is deep into making strong cases for particular teams that have a chance at winning the national title. One argument that is often heard goes something along the lines of “XYZ team has all of the momentum right now and is the hottest team in college basketball. They are the LAST team I’d want to face in the tournament.” With such a strong emphasis on how a team finishes a season as opposed to what they accomplished in the beginning, early predictions and rankings are often overlooked and only brought back as an afterthought for an easy laugh back at the time you thought THAT team was going to win it all. Yet year after year, thousands of analysts do their best to create a hierarchy of the college basketball landscape. Are they just wasting their breath, adding filler to increase anticipation for the new season? Or are we actually much better at predicting future success than we often give ourselves credit for?
A trip down memory lane
To answer this question, I chose to look at the success of early title challengers in the NCAA tournament. A naïve investigation would entail checking the teams in the preseason rankings and comparing with the eventual national champion. If the winner is on the early list, then success! Otherwise, failure. This seemed too simplistic and destined for disappointment, so I opted instead to cast a wider net to view the representation of the preseason Associated Press Top 25 rankings in the NCAA tournament.
To accomplish this, I first gathered 33 years worth of NCAA tournament data ranging from 1985 to 2017. I decided to ignore the first round/play-in games (depending on what they decided to call it that year) and begin my investigation with only 64 teams. Until recently, the teams playing in those preliminary games tended to be smaller schools that won an automatic bid from a lesser-known conference and generally aren’t anywhere near the top 25 throughout the season. Starting from the round of 64, I simply took a tally of the number of teams in each round that were also ranked in the AP’s preseason top 25 list and found the average number of top 25 teams that participated in a given round, except in the case of the rounds of 64 and 32 where I divided by 25 to reflect the maximum possible number of teams that could be involved.
As the numbers above show, approximately 20 teams listed in the preseason rankings each year end up making the tournament, and nearly 80% of the teams in the final four tend to be early favorites. Further, the eventual champion can be found amongst the preseason’s best 85% of the time.
Perhaps counter-intuitively, the Round of 32 as well as the Sweet 16 prove to have the weakest correlation between early forecasts and actual representation with only 61% of possible teams making it to each of those rounds. Many would argue that results like these make the tournament exciting by providing early-round upsets of teams that get caught up in the moment, or horribly underestimate their opponents. As the field dwindles and the initial chaos wanes, the heavyweights begin to re-establish themselves as the dominant teams, driving up the survival rate of preseason top 25'ers higher.
Hail to the new king(s)!
What are some takeaways from this information? First, it’s safe to say that, especially more recently, the teams that are ranked early on tend to find lots of success in the postseason. In fact, the rolling mean shows overall continued improvement in the representation of preseason top 25 teams in the NCAA tournament.
As we can see from the chart above, there is a general trend that teams in the first rankings of the season have performed better with the passing of time. Perhaps this is a product of the perennial top contenders (eg. Kansas, Duke, Kentucky, UNC, Michigan State, …) recruiting increasingly stronger freshmen classes, forcing a further divide between “the best” and “the rest.” Perhaps college basketball is just getting easier to predict. Regardless of the reason, our early-season forecasts are becoming better at predicting which teams will dominate throughout the season.
As the preseason rankings get released each year, don’t be too quick to disregard it as useless information as you are looking at approximately a third of that year’s NCAA tournament field. On the other hand, don’t get your hopes up on your team making a magical Final Four run if they aren’t on the list, and you might as well forget about any dreams of seeing them cut the net in early April. Though it’s called March Madness for a reason, there’s also a very good reason that the early favorites tend to be familiar each year.