A Beginner’s Guide to Understanding Advanced Statistics: Corsi and Fenwick

Max Pacioretty has been a positive Corsi player in 6 of his 8 NHL seasons; what does that mean? – Photo courtesy of Canadian Press

Shot metrics have taken the NHL by storm this season. It’s about time All About The Habs provided a simplified version of the most predominant statistics. This will serve as a reference point for those looking to begin utilizing these metrics.

 

What is Corsi?

It is a plus/minus rating or percentage obtained by simple addition, subtraction and/or division of shot attempts; also referred to as on-ice events. This includes shots on goal, shots that miss the net, shots the opposition blocks and of course, goals. Corsi can be separated by specific on-ice situations – such as even strength as a whole, 5 on 5 play, 4 on 4 play, special teams – or as an overall total including all situations. Corsi can also be used at the player level rather than team level.

The method in calculating the percentage vs. the plus/minus is slightly different.

To determine the Corsi +/-, take the positive shot attempts and subtract the negative shot attempts.  For example:

PaciorettyCorsi

While Max Pacioretty was on-ice, his team registered 11 shots on goal, 3 missed shots and 6 blocked shots.
While Max Pacioretty was on-ice, the opposition registered 4 shots on goal, 2 missed shots and 3 blocked shots.
(11 shots + 3 missed + 6 blocked = 20 shot attempts for) – (4 shots + 2 missed + 3 blocked = 9 shot attempts against)
Max Pacioretty’s Corsi +/- for the game was +11 (20 for – 9 against).

To determine the Corsi%, take the positive shot attempts and divide by the total number of shot attempts by both teams. The outcome will be the percentage of positive shot attempts. Using the above data for the following example:

Max Pacioretty was on-ice for 20 shot attempts for and 9 shot attempts against, meaning there were a total of 29 events.
Max Pacioretty’s Corsi% for the game was 68.9% (20 for / 29 total).

The calculations are identical at a team level. For example:

MontrealBostonCorsi

 

Montreal Canadiens managed 29 shots on goal, 10 missed shots, and 18 blocked shots. If you add these events together, you get 57 Corsi events for the Canadiens (29 + 10 + 18).
Boston Bruins managed 19 shots on goal, 6 missed shots, and 15 blocked shots. If you add these events together, you get 40 Corsi events for the Bruins (19 + 6 + 15).

As a +/- rating:
Montreal was +17 on the night in Corsi events (57 – 40).
Boston was -17 on the night in Corsi events (40 – 57).

As a percentage:
Montreal’s Corsi% on the night was 58.7% (57 for / 97 total).
Boston’s Corsi% on the night was 41.2% (40 for / 97 total).

 

What does Corsi reveal?

Corsi is commonly used as a time of possession surrogate. While it doesn’t quantify the exact duration a team spends in the offensive or defensive zone, it does provide a strong notion of how often a team entered the offensive zone or was forced into the defensive zone. This can go a long way in determining who out-played whom.

The objective of a hockey game is to score more goals than your opposition. The best way to accomplish this is by registering shots on goal; more specifically shot attempts because it’s impossibly inaccurate to guarantee every shot hits the net.  This is where Corsi peeks its head. Statistically speaking, the more shot attempts you manage, the more likely you are to score.

 

Any Corsi No-Nos?

First and foremost, never draw conclusions from a single advanced metric. They must be used in combination with others and context is

(Minas Panagiotakis/Icon SMI)

(Minas Panagiotakis/Icon SMI)

just as important as the statistic itself. The role a player fills largely affects Corsi. For example, if Michel Therrien deployed Lars Eller primarily against John Tavares in the defensive zone, it would be much more difficult for Eller and his linemates to generate shot attempts than it would be for David Desharnais who is primarily deployed against Brandon Sutter in the offensive zone. Ergo, Eller’s Corsi% would reflect that in a very negative way.

Finally, be weary of the term “sample size”. Any type of analysis using a small portion of games will be immediately shot down. This is due to inevitable fluctuation in data in smaller samples. If possible, do your best to use a sufficient sample size of games or time-on-ice.

 

What is Fenwick?

Fenwick is Corsi minus blocked shots. In other words, it’s derived by using goals, shots on goal and shots that miss the net. Like Corsi, Fenwick can be separated by each on-ice situation and can be used at both an individual and team level, as well as being represented as either a percentage or a plus/minus rating.

The reason Fenwick excludes blocked shots is because there are plenty of thoughts about blocking shots being an art and a talent. It wouldn’t make sense to penalize a team for exhibiting a talent they’ve worked to develop.

The method in calculating Fenwick% and Fenwick +/- is identical to that of Corsi.

To determine the Fenwick +/-, take the positive shot attempts and subtract the negative shot attempts.  For example:

PaciorettyFenwick

 

While Max Pacioretty was on-ice, his team registered 11 shots on goal and 6 missed shots.
While Max Pacioretty was on-ice, the opposition registered 9 shots on goal and 2 missed shots.
(11 shots + 6 missed = 17 shot attempts for) – (9 shots + 2 missed =11 shot attempts against)
Max Pacioretty’s Fenwick +/- for the game was +6 (17 for – 11 against).

To determine the Fenwick%, take the positive shot attempts and divide by the total number of shot attempts by both teams, remembering to exclude the blocked shots. The outcome will be the percentage of positive shot attempts. Using the above data for the following example:

Max Pacioretty was on-ice for 17 shot attempts for and 11 shot attempts against, meaning there were a total of 28 events.
Max Pacioretty’s Fenwick% for the game was 60.7% (17 for / 28 total).

 

What does Fenwick reveal?

Like Corsi, Fenwick is a nice tool to help determine which player or team possesses the puck more. Fenwick, however, has a greater correlation with scoring chances than Corsi does but less of a correlation with time of possession due to the fewer events as a result of

(Photo: John Russell/NHLI)

(Photo: John Russell/NHLI)

excluding blocked shots. The predictive value Fenwick carries is represented nicely by former Eyes On The Prize writer Chris Boyle here.

Players who block a high volume of shots tend to be very poor Corsi players and much better Fenwick players. As a result, a weak Corsi% doesn’t always mean that player is completely ineffective. Defensemen such as Kris Russell, Josh Gorges and Roman Josi aren’t the best possession players, but due to the volume of shots they block, their Fenwick numbers look slightly more respectable.

 

Any Fenwick No-Nos?

Due to both Corsi and Fenwick being possession metrics derived from shot attempts, the mistakes for both are essentially identical.

 

Final Thoughts

These are predominately the most mainstream advanced statistics being tracked to date. Corsi and Fenwick data are used as starting points for many other metrics such as Quality of Competition/Teammates, Relative Statistics, Individual Shot Attempts, and Scoring Chance %, among others. In combination with other metrics, Corsi and Fenwick can drastically aid in evaluating player and team talent.

 

Share Button

Leave a Reply

Your email address will not be published. Required fields are marked *

1 × 3 =

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Habs Headlines (Podcast): Season 2, Episode 1

  MINAS PANAGIOTAKIS / GETTY IMAGES   HabsHeadlines is making a comeback. Patrick Tallon and Drew Hardiman recorded their first podcast together following the Habs 3-2 win over the Chicago Blackhawks Sunday night. From covering the lineup decisions, prospects, coaching staff, Marc Bergevin, and taking questions from the mailbag, you can be assured that no […]

Share Button

JOIN THE CONVERSATION