Fenwick (FF%)
A shot attempt metric that counts unblocked shots (shots on goal plus missed shots), excluding blocked shots. Considered slightly more predictive of future outcomes than Corsi.
Fenwick (named after blogger Matt Fenwick) is a close cousin of Corsi that excludes blocked shots from the count. The rationale is that blocked shots introduce noise: a team might have a high shot-blocking rate as a deliberate strategy, or blocked shots may not truly represent offensive pressure the way shots on goal and missed shots do.
By removing blocked shots, Fenwick provides a slightly cleaner signal of which team is generating and allowing genuine scoring chances. Research has shown Fenwick to be marginally more predictive of future goal scoring than Corsi, though the difference is small. FF% (Fenwick For Percentage) above 50% indicates the player's team is generating more unblocked shot attempts than it allows.
In modern analytics, both Corsi and Fenwick have been somewhat superseded by expected goals (xG) models, which weight each shot attempt by its probability of becoming a goal. However, Fenwick remains valuable as a quick, transparent possession measure that does not require a complex model.
Formula
FF (Fenwick For) = Shots on Goal + Missed Shots (for) FA (Fenwick Against) = Shots on Goal + Missed Shots (against) FF% = FF / (FF + FA) x 100
Frequently Asked Questions
How is Fenwick different from Corsi?
The only difference is that Fenwick excludes blocked shots. Corsi counts all shot attempts (on goal + missed + blocked), while Fenwick counts only unblocked attempts (on goal + missed). They correlate very highly but Fenwick is considered slightly more predictive.
What is a good Fenwick percentage?
Similar to Corsi, an FF% above 50% is positive. Elite possession players sustain 54-57% FF% at 5v5. The scale and interpretation are essentially the same as CF%.
Why would blocked shots be excluded?
Blocked shots can reward a defensive strategy (some teams deliberately funnel shots to areas where defenders can block them) and may not truly represent offensive pressure. Removing them provides a cleaner measure of genuine scoring chance generation.
Is Fenwick still relevant with xG models available?
Yes. Fenwick is transparent, easy to calculate, and requires no model assumptions. It remains useful for quick analysis and as an input feature in more advanced models. However, for evaluating individual players, xG-based metrics provide a more complete picture.