In a sports world driven by data analytics, Noah Seaner ‘25 has been at the forefront of it all for the men’s basketball team. When he’s not recording game statistics, however, he’s busy running tens of thousands of simulations to see how the season will play out.
“It’s stuff I’ve done before, but I’ve never really put it into a real-world scenario,” Seaner said.
Alongside his current role as Lafayette College’s manager of basketball coaching analytics, Seaner has spent the entire season fine-tuning his own side project, which predicts the final record for all Patriot League teams based on how they have performed up to each simulation date.
The program, built in R — a programming language used for statistical analysis — pulls in each team’s schedule, average possessions per game and offensive and defensive efficiency to estimate expected possessions and scoring rates for each conference game. Seaner first ran the model after teams finished their non-conference schedules, and now updates it on a game-by-game basis.
Seaner’s predictions have remained consistently accurate throughout the season. His 15-game simulation update projected the Naval Academy, Colgate University and American University as the top-three regular season finishers. All three teams were initially predicted to top the standings in his preseason report, all of which sit in the top four heading into the last game of the season. Additionally, both Holy Cross and Bucknell University have remained in the bottom three all season in both his predictions and in the conference standings.
“The results of that one were pretty similar to the ones that we’re seeing now,” Seaner said. “Usually, the teams at the top and the teams at the bottom have stayed pretty similar.”
Seaner explained that he used repeated random sampling to obtain a wide range of numerical outcomes — to simulate hypothetical head-to-head matchups 10,000 times.
“The most difficult part was just getting the simulation to run once,” he said. “Once I could get it to run once, then you can get it to run as many iterations as you want. The code itself is a lot more intensive than what I am used to.”
In his time at Lafayette, Seaner was responsible for logging box score statistics and analyzing game film for the Leopards, which would be incorporated into projects that provided additional insight for the team’s coaches. After graduating last May, however, Seaner has stayed in a “transitional role” in which he oversees current Lafayette students who collect data for the team.
“I told him, why don’t you spearhead everything with this and take some basic managerial control over the analytics team?” said assistant coach Nikolai Arnold, who initially established the data analytics program ahead of last season. “He’s done a phenomenal job of showing to be able to handle the day-to-day. His ability to be creative and think of things that are outside of the box has been super valuable to us.”
Among the projects Seaner has spearheaded this season is a predictive model that calculates the team’s expected points in a given game based on the shots taken. Though the record simulation was intended to serve as a side project, Arnold said the coaching staff takes “serious value” in the tool.
“As far as trying to adjust based on what we’re doing well and what we’re not doing well, and hopefully trying to get better at the things that will help us win more games and flip those standings around,” he said.
Seaner’s stretch goals for the project include expanding the simulations to predict results for both the Patriot League and the NCAA Tournament.
“Once our conference season’s over, I will revisit this and resim it just to see if Lafayette has a chance,” Seaner said. “I’m also thinking about applying this for the overall March Madness tournament, which might be a little bit more intensive.”