This study aims to evaluate the predictive power of college basketball statistics for success in the NBA, using players selected in the drafts between the 2010 and 2017 seasons. The goal is to identify the variables that contribute positively or negatively to an athlete’s Value Over Replacement Player (VORP), aiming to establish a relationship that assists in identifying selection preferences for the NBA draft. To analyze the dataset, dimensionality reduction techniques such as Principal Component Analysis (PCA) were applied to evaluate the predictive capacity of the selected variables. Multiple Linear Regression and Artificial Neural Networks were also used for VORP prediction. The results show that, although it is not possible to distinguish a new star using college statistics alone, machine learning techniques serve as an auxiliary tool to improve draft ranking.