60 seconds. The undivided attention of a childhood hero. A half decade of tireless work yearning for his approval.
RCS intern Shael Brown found himself in a once-in-a-lifetime moment as he stared into the impatient eyes of Toronto Raptors president Masai Ujiri, his back to a crowd of eager attendees at the MIT Sloan Sports Analytics Conference.
A Nova Scotia native and lifelong Raptors fan, Brown had just one minute to convince Ujiri that he had developed a new statistic that would help the Raptors front office and coaches better understand the impact their players had on each game – and that those 60 seconds would be well worth the busy Ujiri’s time.
The irony? Brown’s statistic – whose accompanying research paper earned him a prestigious spot presenting his work at the Sloan Conference, the premier gathering of the wisest minds in sports and statistics – was initially not even impressive enough to convince his high school teacher to give Brown a passing grade on the project.
Inspired by the movie Moneyball, a teenaged Brown fell in love with the idea of combining his passions at school – math – and away from it – sports. However his initial attempt at developing his very own statistic fell short; when he submitted his statistic for a school assignment, he received a failing mark.
Years later, after collecting a high volume of data and adjusting the scaling of the metric, Brown unveiled a new and improved version of the statistic – which he calls the Integrated Playmaking Metric (IPM). IPM gives more meaning to a part of the game that is often difficult to derive value from: passing.
In free-flowing passing sports such as basketball, soccer, and hockey, the ball is constantly being passed from one player to another, yet the only real statistic used to measure the effectiveness of a player’s passing is the assist. Singling out the one player who had the final pass of a scoring possession is a flawed strategy, however, as it ignores the importance of build-up play – an essentially ingredient to every goal-scoring opportunity. IPM resolves this issue by assigning a score to each player based on the “value” of their passes.
For instance, a player who constantly passes balls back to his defenders in soccer adds less to his team’s attack than a player who consistently moves the ball to the feet of the team’s most effective attackers. Getting the ball in possession of the players who make things happen is just as important as those players making things happen.
The Cliff’s Notes version of how the statistic is calculated is only slightly less complex than the unabridged edition. Essentially, IPM utilizes the same algorithm that Google uses to determine which websites are the most important based on a specific search. Instead of webpages, however, this Google Page Rank algorithm is applied to basketball, soccer, and hockey players to determine which passers are the most important.
Bolstered by additional data and a tweaked metric, the Sloan Conference noticed the value of Brown’s metric where his high school teacher did not, and offered him a spot to present a poster of his research at the conference.
It was here, in the halls of MIT, that Brown’s presentation caught the attention of RCS’ Head of Business Development, Jeff Heimbold. Two months later, Brown found himself at the Reality Check headquarters in Burbank, CA, integrating his IPM statistic into RCS’ LaunchPad, a touchscreen tool for in-depth soccer analysis.
Using the large volume of data from OPTA already present on LaunchPad, Brown helped add the IPM statistic to LaunchPad’s suite of “Match Analysis” tools, but he didn’t stop there.
“I had this fun idea where we have all this information about where shots were taken,” said Brown. “Anytime where you have large amounts of spatial data like x,y coordinates, a natural thing to do is look at clusters and essentially ask, ‘Where were the majority of these kind of things?’”
The answer to that query provided Brown and RCS with yet another exciting tool for data visualization on LaunchPad. Looking at the areas in which a team had a dense cluster of shot attempts overlaid on a similar shot plot that detailed where the opposing team’s defense allowed shots to be taken, Brown determined that these overlapping clusters – or “danger areas” – were spots on the pitch where he could predict a goal might come from.
Testing his analysis in the Champions League semifinal between Atlético Madrid and Real Madrid, Brown found that his pre-match prediction of danger areas held up, adding yet another analysis tool to the future of LaunchPad.
Despite spending just one month with RCS, Brown said his experience with the company taught him integral lessons that will aide his future as a statistician – primarily the importance of data visualization.
“While maybe I could look at an unscaled statistic and be able to say, ‘Ah yes, this mean x y z,’ that’s not for anyone that doesn’t have a math or stats background,” Brown said.
Although finding patterns in the data or developing metrics to predict or analyze a result is useful, it loses its value if the general public cannot easily understand, which is what makes the marriage of Brown’s statistic and LaunchPad’s visualization such a perfect union.
After completing his RCS internship and returning home to Canada, Brown’s schedule doesn’t calm down. He has another internship lined up focusing specifically on data analytics to hone his skills in that field, then will head to graduate school in Montreal where he plans to continue conducting more research while hoping to work part-time for a hockey data analysis company.
But Brown’s eyes haven’t strayed far from his initial target. After completing his graduate degree, his goal is to once again submit his research to the Sloan Conference with the hopes of one day securing a front office data analysis job with an NBA team. But this time around, armed with real life experience, a graduate degree, and a more widely-used IPM statistic, perhaps it will be Masai Ujiri pleading for just 60 seconds of Brown’s time.