Abstract
Slot operations are crucial to the success of U.S. casinos, yet little is known about the variation in unit‐level performance. This article combines the statistical methods of Principal Components Regression (PCR), population partitioning, and a method from Computational Geometry (Voronoi Diagrams) in slot performance‐potential research. This approach is advanced in an attempt to improve the prediction process from existing methods. The results indicate success in this regard. The final output is a map of the casino floor, showing where the underperforming and overperforming units are located. When used in concert with performance data, this map is especially helpful to casino executives, allowing them to identify and profile overperforming and underperforming areas of the slot floor. Going forward, such findings offer valuable casino design insight for operators and developers alike.