Prediction With Conviction: An Application of Relevance-Based Prediction to the NBA

This Version September 17, 2025

Megan Czasonis is a Founding Partner of Cambridge Sports Analytics in Cambridge, MA.

100 Main Street, Cambridge MA, 02142 mczasonis@csanalytics.io 

Mark Kritzman is a Founding Partner of Cambridge Sports Analytics in Cambridge, MA, and a senior lecturer at the MIT Sloan School of Management in Cambridge, MA. 

100 Main Street, Cambridge MA, 02142 mkritzman@csanalytics.io 

Cel Kulasekaran is a Founding Partner of Cambridge Sports Analytics in Cambridge, MA.

100 Main Street, Cambridge MA, 02142 ckulasekaran@csanalytics.io 

David Turkington is a Founding Partner of Cambridge Sports Analytics in Cambridge, MA.

100 Main Street, Cambridge MA, 02142 dturkington@csanalytics.io 

Abstract

The authors describe a new prediction system, called relevance-based prediction (RBP), for predicting player performance for NBA draft prospects based on the outcomes of previous NBA players. This approach rests on a statistical concept called relevance, which gives a mathematically precise and theoretically justified measure of the importance of a previous player to a prediction. The authors also describe fit, which gives advance guidance about the reliability of a specific prediction. And they show how fit, together with asymmetry, focuses each prediction on the combinations of predictive variables and previous players that are most effective for that prediction task. The authors argue that RBP addresses complexities that are beyond the reach of conventional prediction models, but in a way that is more transparent, more flexible, and more theoretically justified than widely used machine learning algorithms. 

Read the full paper on SSRN

This research was authored by Megan Czasonis, Mark Kritzman, Cel Kulasekaran, and David Turkington.

Published as an MIT Sloan Research Paper (No. 6955-23) and available via SSRN.