How It Works
The model
behind it.
Sprint Calculator combines two complementary techniques. For single-input cases — such as converting a hand-timed mark to an electronic equivalent — we use simple linear regression, optimised for precision on sprint-specific data.
For multi-input predictions, we use Extreme Gradient Boosting (XGBoost), trained on a database of approximately 10,000 official 100m and 200m races. The result accurately distinguishes all types of sprinters and analyses individual performance trends.
@sprintcalculator