Corsi and Fenwick Explained for NHL Bettors

Advanced stats changed how hockey teams evaluate talent, and they should change how you evaluate bets. Corsi and Fenwick are the two most widely cited possession metrics in hockey analytics, and they measure something the box score ignores entirely: which team is controlling the play. Goals are noisy — a team can dominate for 58 minutes and lose on a fluke deflection. Corsi and Fenwick cut through that noise by counting the attempts, not just the results. For bettors, these metrics are a window into what is actually happening on the ice beneath the final score.

See also expected goals xG in NHL betting.

What Is Corsi

Corsi measures shot attempts. It counts every shot directed at the net — shots on goal, missed shots, and blocked shots — at even strength. If a team generates 60 Corsi events for and allows 45 against during a game, their Corsi For percentage (CF%) is 57.1%. That means they directed 57.1% of all shot attempts while they were at even strength, which is a strong indicator of territorial dominance.

The metric was named after Jim Corsi, a former NHL goaltending coach, though its application as an analytics tool was popularized by bloggers and analysts in the late 2000s. The core idea is simple: the team that generates more shot attempts is generally the team that controls the puck more, spends more time in the offensive zone, and creates more opportunities to score. Over a large enough sample, this control translates into goals and wins.

Corsi is typically expressed in three ways. CF% (Corsi For percentage) is the most common, showing the share of total shot attempts a team generates. CF/60 (Corsi For per 60 minutes) measures raw volume of shot attempts per hour of even-strength play. Corsi Rel compares a player’s or line’s CF% to the team’s CF% when that player or line is off the ice, isolating individual impact from team effects.

A CF% above 52% is generally considered strong. A CF% above 55% is elite and suggests a team is dominating possession to a degree that is difficult to sustain without genuine systemic advantages. Below 48%, a team is being badly outplayed territorially, even if the scoreboard does not always reflect it.

What Is Fenwick

Fenwick is a refinement of Corsi that removes blocked shots from the equation. It counts only shots on goal and missed shots, excluding any attempt that was blocked by an opposing player before reaching the net. Fenwick For percentage (FF%) is calculated the same way as CF% — Fenwick events for divided by total Fenwick events — and it is interpreted similarly.

The rationale behind Fenwick is that blocked shots introduce noise. A blocked shot is partly a function of the shooting team’s attempt and partly a function of the defending team’s positioning and willingness to block. By excluding blocks, Fenwick isolates the shooting team’s ability to generate unimpeded attempts, which some analysts consider a cleaner measure of offensive generation.

In practice, Corsi and Fenwick are highly correlated. Teams that rank well in CF% almost always rank well in FF%, and the differences between the two metrics are usually small — a percentage point or two. The situations where they diverge meaningfully are rare but informative: a team with a much higher CF% than FF% is generating a lot of shot attempts that get blocked, which may indicate a reliance on low-percentage shots from the perimeter that defenders easily step in front of.

Context Matters: Score Effects and Zone Starts

Raw Corsi and Fenwick numbers are useful but incomplete without context. The two most important adjustments are score effects and zone starts.

Score effects refer to how a team’s shot attempt numbers change based on the game state. A team that is trailing by a goal tends to generate more shot attempts because they push forward aggressively, while the leading team sits back and protects its lead. This means a team that trails frequently might post a decent raw CF% despite being a weaker team overall, simply because they spend more time in desperation mode. Score-adjusted Corsi (often labeled SCF% or score-close CF%) filters out this distortion by looking only at even-strength play when the score is within one goal, which produces a more accurate picture of territorial dominance.

Zone starts measure where a player or team begins their even-strength shifts. A player who starts 65% of his shifts in the offensive zone will naturally post a higher CF% than one who starts 65% in the defensive zone, because he begins each shift closer to the opponent’s net. Adjusting for zone starts levels the playing field and prevents coaches’ deployment decisions from skewing the numbers. When evaluating Corsi or Fenwick for betting purposes, always look for zone-start-adjusted figures — they tell you what a team or player does with neutral circumstances rather than favorable ones.

Using Corsi and Fenwick in Betting

The practical application for bettors is straightforward: Corsi and Fenwick help you identify teams whose results do not match their underlying process. A team with a strong CF% but a poor record is probably getting unlucky — their goaltender is underperforming, their shooting percentage is abnormally low, or they are losing a disproportionate share of one-goal games. History shows that these teams tend to improve over time as luck normalizes, and their current odds may reflect their poor record rather than their true quality.

The reverse is equally valuable. A team posting a 55% winning rate with a 47% CF% is almost certainly overperforming. They are winning despite being outplayed territorially, which usually means they are riding hot goaltending, an unsustainably high shooting percentage, or both. These teams are prime regression candidates, and betting against them — particularly at inflated odds driven by their misleading record — can be a profitable long-term strategy.

For totals betting, Corsi provides insight into game pace. A matchup between two high-CF% teams is likely to feature a high number of shot attempts in both directions, which increases the probability of more goals and favors the over. A matchup between two low-CF% teams — clubs that play a conservative, low-event style — will likely produce fewer attempts and fewer goals, leaning toward the under. Cross-referencing team-level CF% with the posted total can flag games where the book has set the number too high or too low relative to the expected shot volume.

On a player level, Corsi Rel is useful for prop betting. A forward with a strong Corsi Rel is driving play when he is on the ice, generating more attempts for his team and suppressing attempts against. This player is more likely to be involved in scoring plays, which makes his goal, assist, and shot props more attractive. Conversely, a player with a poor Corsi Rel is a drag on his team’s possession, and his prop lines may overestimate his offensive output.

Limitations of Corsi and Fenwick

Corsi and Fenwick are powerful tools, but they are not the whole picture. The most important limitation is that they measure quantity, not quality. A team that fires 35 shots from the perimeter generates the same Corsi as a team that fires 35 shots from the slot, but the second team is creating far more dangerous chances. This is why expected goals (xG) models have become increasingly popular — they weight shot attempts by location and type, providing a quality-adjusted view that Corsi and Fenwick cannot.

Another limitation is sample size. Corsi stabilizes over roughly 20-25 games for a team and longer for individual players. Using a five-game Corsi sample to evaluate a team is like judging a hitter after 15 at-bats — the noise overwhelms the signal. Bettors who react to short-term Corsi swings without waiting for the data to stabilize are chasing variance, not skill.

Special teams also fall outside the Corsi and Fenwick framework. Both metrics are measured at even strength only, which means they tell you nothing about a team’s power play or penalty kill. A team might dominate at five-on-five but bleed goals on the penalty kill, and Corsi would not flag that vulnerability. For a complete assessment, you need to combine even-strength possession metrics with special teams data and goaltending numbers.

Finally, Corsi and Fenwick do not account for game management. Some coaches deliberately concede territorial control to opponents, preferring to play a counterattacking style that absorbs pressure and strikes on the rush. These teams post low CF% numbers by design, not by deficiency, and they can be profitable bets if the market penalizes them for what looks like poor possession but is actually a coherent strategic choice.

Beyond the Shot Counter

Corsi and Fenwick were revolutionary when they entered the hockey analytics conversation over a decade ago. They gave fans, analysts, and bettors a way to measure what the eye test could only approximate: who was actually controlling the game. That contribution remains valuable, even as more sophisticated models have emerged.

For bettors, the lesson is not to worship any single metric but to use Corsi and Fenwick as one layer in a multi-layered analysis. Check the CF% to understand territorial control. Check score-adjusted numbers to remove the noise of game state. Then layer in shot quality, goaltending, special teams, and schedule context to build a complete picture. The bettor who integrates possession data into a broader framework will consistently outperform the one who ignores it — and the one who relies on it exclusively. The truth, as usual, lives somewhere in the middle.

Corsi and Fenwick for NHL bettors at hockey betting.