Optimizer + Corsica Machine Learning Projections

The use of state-of-the-art machine learning methods and access to the rich Corsica database have allowed us to significantly improve our Daily Fantasy projections.

Through exhaustive feature engineering and analysis, we identified some of the most potent predictors of Fantasy Points for both skaters and goaltenders. In addition to historical player-specific performance, a multitude of team statistics including Expected Goals are used, helping to identify environments conducive to extreme DFS output.

Our models were trained on hundreds of thousands of player-games and meticulously tuned to squeeze the very most out of the data. We employed an extremely powerful gradient boosting algorithm to effectively learn how to better learn. The result is a 12.5% improvement over our former projections.

Both Cash and GPP projections are offered. Cash projections indicate what a player's average performance might look like, while GPP reflects a top-end performance. A player's fantasy output should have a roughly 5% chance of equaling or exceeding the GPP projection.

*Root Mean Squared Error (or RMSE) is a measure of difference between two sets of values. A lower score means predicted values are more accurate.

RMSE (Fanduel)RMSE (Draftkings)RMSE (Yahoo)
12.5% better13.3% better11.7% better