Carl Colglazier


Cheesecake is an evidence-based scouting and statistics approach to the FIRST Robotics Competition. It has two main goals:

  1. To use existing data on the results of FIRST competitions to create measurably accurate prediction metrics.
  2. To facilitate a mixed-methods (quantitative plus qualitative) approach to FRC scouting specific to each game.

Cheesecake takes an empirical, evidence-based approach to how it handles FRC data.


Scouting is a large commitment for a team. At most competitions we attend, we usually allocate a significant amount of team resources to ensure we have as much data as possible on each robot at the competition. This information typically goes into a binder and is used by our scouting team to determine the best robots to pick during the alliance selections and the optimal strategies to play against individual robots.

The goal of Cheesecake is to ensure that scouting information is transformed into useful metrics. It draws inspiration (and some models) from other types of sports analytics, statistics, and previous related systems.