Fancy Stat Summer School - Corsi
Over the summer, EOTP is going to break down every fancy stat in order to help people understand what we're talking about all the time. Hopefully this series serves as a frame of reference throughout the future of the site.
What is Corsi?
Corsi is an inherently simple concept. The number is derived by shot attempts; meaning goals, shots on goal, shots that miss the net, and shots that are blocked. It is usually represented as either a percentage, or as a plus/minus rating.
Corsi can be used at both the team or individual level and is calculated by situation, meaning separate numbers for even strength and special teams.
When used as a percentage, it is calculated by dividing positive shot attempts by total shot attempts between both teams, the resulting percentage representing the proportion of positive shot attempts in a given situation.
When used in plus/minus format, it is calculated by subtracting negative shot attempts from positive ones.
While putting Corsi in a percentage format gives a clearer picture of a player or team's performance, the plus/minus format can add additional context if a player or team is higher event than another.
As with all possession metrics, Corsi can also be broken up into for and against statistics, which is a better usage if you want to specifically measure specifically offense or defense.
Why is it called Corsi?
Former Buffalo Sabres goaltending coach Jim Corsi is the one who created the statistic, which was initially intended to be a measure of how much work a goaltender is subjected to each game, as a goaltender has to react to every shot attempt, not just every shot. However the statistic ended up being more useful to measure the performance of skaters instead of goaltenders.
What does Corsi tell us?
Corsi is most appropriately used as a proxy for time of possession. The reason it is used this way is that the goal of the game of hockey is to score goals. In order to score goals you have to generate shots, which means shot attempts. Statistically, the more shot attempts there are, the more likely you are to get a goal.
Being statistically more likely to get a goal when you have more shot attempts than your opponents means that Corsi has predictive power in future games. A strong Corsi possession team is likely to score more goals than a weak Corsi possession team. However of the methods designed for predicting future goals, it does not have the highest correlation.
For this reason I find that Corsi is best used as a metric to measure time of possession. Time of possession is important for scoring, but also important for defense. It may seem obvious, but if you have the puck, the other team can't score. Even a player who has minimal offensive value can play an important role if he outpossesses the other team.
What are Corsi's limitations?
Corsi can not be used on it's own to evaluate an individual player. There is more parity in the NHL than ever, but a player who is hard matched to take on Sidney Crosby while starting most of his shifts in the defensive zone will have a much tougher time generating shot attempts than someone who is matched against a fourth line and starts their shifts exclusively in the offensive zone. The role a player fills has a large impact on their Corsi numbers.
As with all statistics, there are some who overestimate their application. At EOTP, regulars will know that we treat player analysis the same way a doctor will treat antibiotic resistent bacterial infection; with a cocktail. Usually if you're using a single statistic to evaluate a player, without context, you're going to do a poor job.
There are some who say that making decisions based on Corsi numbers from a player personnel point of view would be a bad idea because players would learn to game the system and just shoot from everywhere to boost their individual ratings, although this seems to ignore the fact that a coaching staff would notice this and likely bench the player.
The second to last game against Buffalo that the Canadiens played in 2013 was one of the most dominant even strength performances of the season. As you can see in the table below, every single Habs player had a positive differential on the night. Including the goaltender allows us to have a team differential as well. For each category, the left side is "for" and the right side is "against". If you use the previously laid out formula, you can calculate the Corsi percentage for every player on the roster with the information available.