Quantifying Finishing Skill in Football



Quantifying Finishing Skill - Statsbomb

Does anyone see a reason this couldn’t be implemented for hockey? I know it is frequently mentioned over the course of a season that a player’s shooting percentage is elevated or depressed compared to league or career averages.

If finishing skill were indeed quantified, it could help identify if that variance is likely correct, and potentially to what degree. Furthermore, couldn’t we estimate the xG lift from these players taking a shot if finishing skill was quantifiable?


I recall seeing the same methodology used to estimate NHL goaltenders’ true save %. as for shooters, it’s definitely possible to do the same but the main issue is variance - notice how the confidence intervals are so big for the soccer players that you cant come to any clear conclusions.


I believe it’s similar to cole anderson’s approach for xG lift for goaltenders true save %. I’m wondering if the shot volumes being considerably higher in hockey would help lower the variance