Cross validation output for forwards in the 2020 NHL draft (version 2020.2.0)
This is cross-validation data. This is the result of applying the model for the 2020 draft on the data used to train the model.
See the performance data at the end of the page.
A player can be missing if certain aspects of the overall dataset are not rich enough to make a prediction. For example, the player may not have played many games in a single league, or may play in a more obscure league.
Projected NHL career data is extrapolated, but based on
real to-date statistics. Thus, projected data is only available for past drafts, which are used to evaluate the model.
Predicted NHL data is data that was predicted using only pre-draft information.
A player's predicted values are based on a player's probability scores for the given classes.
Projected NHL P per 60Predicted NHL P per 60Miss Prob.4th-Liner Prob.Solid Prob.First-Liner Prob.Actual Overall DraftDraft Year
Reported below are the area-under-the-curve (AUC) scores for the class probabilities shown above. This score is a measure of the reliability or accuracy of the predicted probabilities. 1.0 is perfect, and 0.5 is no better than a coin-flip. See this Wikipedia article for more information.