So, two-time reigning NBA MVP Stephen Curry has picked the Blue Devils to win it all this year because they can “get hot at the right time and they’re gonna stay hot.” Not to mention, his brother Seth used to play for Duke. How about you? Have you filled out your own bracket? How can you predict how your picks will fare?
While the chances of picking the perfect bracket are one in 9,223,372,036,854,775,808, you have a fifty-fifty chance of predicting who emerges in each matchup. Using historical data (and barring tournament-changing injuries to key players), you can objectively analyze each matchup in your bracket. Here are some key attributes you can look at in predicting each match up — from the simple, to the complex, and even to the seemingly whimsical (albeit statistically irrelevant):
Tournament Seeding and Ranking
Tournament seeding and AP ranking still remain as a simple, but effective, way of predicting the outcome of each matchup. These attributes reflect how each squad has performed throughout the college basketball season. Thirty-two seeds are given automatic entry by winning their respective conference tournaments. These include top college basketball conferences such as the ACC, A-10, Big 12, Big East, and the Pac-12. The rest of the pack, together with the rankings of the first 32 teams, are determined by a selection committee who considers conference, schedule, win-loss records against other tournament teams, injured players, history of coaches, team performance on the road, and even computer rankings. One can say that the tournament seeding and ranking processes have already done the bulk of the work of measuring the strength of each squad for them.
Still, making predictions based on seeding and ranking alone is not enough. As history would have it, only two top seeds in the past 17 years have won it all.
Tournament appearances reflect how well schools have been performing over the years. Experience is a big factor in any sport. That’s why analysts are careful when betting against schools with coaches and even second or third-year players with experience in the tournament. Tournament appearances also reflect the strength of a school’s basketball program. From 1939 to 2016, seven teams have won the tournament a total of 41 times—a testament to how remarkable their programs are.
There are two important win-loss records that one should consider: the season record and the record against other tournament teams. Like the tournament rankings, these records reflect the performance of each team throughout the year. They also consider their performance when facing elite competition, which may prove quite helpful in predicting how they fare when facing pressure.
A more complex ranking system is the ELO ranking. It was introduced by research and statistics website, FiveThirtyEight, and considers wins and losses, point-differentials, home-court advantages, and even last year’s performance. A critical factor here is the point-differential as it takes the win-loss record to the next level. It measures not only how much each squad is winning, but also in what fashion — whether a team is winning clutch games or demolishing other teams. FiveThirtyEight has done well in predicting the Cleveland Cavaliers as the winner of last year’s NBA championship. Will Villanova fulfill FiveThirtyEight’s prediction of winning it all?
Team Uniforms and Mascots
While they seem statistically irrelevant, it still is difficult to bet against teams donning blue or orange, or those whose mascots are cats or people. Since the NCAA’s expansion 32 years ago, teams wearing blue have won a remarkable 20 times. In terms of winning percentage, teams wearing blue and teams wearing orange are top one and two, respectively, making it difficult to bet against them. When it comes to team mascots, teams with cats or people as mascots have won a combined 12 times, followed by other animals and mythical creatures with five each. (However, all those five championships by mythical creatures belong to Mike Krzyzewski’s Blue Devils.)
Win with Data in Sports and Business
Having more data leads to better informed decisions, whether in sports or business. But oftentimes, the accumulation of large amounts of data can be overwhelming. NCAA made it easier for fans to compare all these relevant (and even seemingly irrelevant) data with its matchup tool. For business, Liaison Technologies helps enterprises integrate their multiple, disparate data sources to make data analysis more effective and efficient. The Liaison ALLOY™ Platform integrates ever-growing numbers of cloud applications and data sources. It focuses on compliance and unification of integration and data management capabilities, and is much better suited to solving today’s (and tomorrow’s) true integration challenges. Contact our data experts to inquire about the Liaison ALLOY™ Platform.