Quarterback Throw

Catapult looks to improve the results
of the initial quarterback detection
algorithm and to quantify the movements
of quarterbacks in a throwing motion.
The development of the new algorithm
included training a machine learning
model on quarterback data from
NCAA Division I quarterbacks and NFL
quarterbacks using Catapult S5 devices.
Testing the algorithm on multiple
quarterbacks yielded the targeted results
above 90% accuracy.
The primary use case of this detection
algorithm is to output PlayerLoad,
RotationLoad and other metrics per throw
for further analysis.