In the world of sports analytics, hockey is relatively new to the game. While many fans hopped on the analytics bandwagon that began in baseball in the 1960s, mainstream sports writers and the NHL did not seem interested in advanced stats. During the 2014-2015 season, the NHL officially partnered with SAP (an enterprise software company) to include advanced statistics on the NHL website.
Most of these advanced statistics go above casual fans’ heads, but these stats can tell fans so much about a player’s overall ability. The problem with hockey analytics is how complex the sport of hockey is. Every decision players make can affect the entirety of the game. There is no magic stat in hockey that can define a player. To understand a player, you must combine all information on them in attempt to predict how they will do the upcoming season. And then, after analyzing your team’s statistics, your favorite player will get traded, the Captain will get injured, and the goalie will forget that they’re actually supposed to stop the puck. It is ridiculous to think that hockey can be predicted with a bunch of formulas. So much can go wrong in so little time. But we can delve into the stats to try to understand just a little better.
One of the most common terms in advanced hockey statistics is Corsi number. This number can be thought of as a better way to measure the shots attempted differential. The Corsi number can be calculated for a team and also a player. For the sake of simplicity, when looking at formulas we are going to find the Corsi number for Toronto Maple Leaf Auston Matthews in a hypothetical game against the Washington Capitals. The formula would look like this:
Matthew’s Corsi Number = (Leafs’ shot attempts while Matthews is on the ice) – (Caps shot attempts with Matthews on the ice)
The less convoluted way to write it is:
Corsi = CF – CA
CF and CA standing for “Corsi for” and “Corsi against”, respectively.
Pretty basic, right? The way CF and CA are found is by adding shots on goal, missed shots on goal, and blocked shots on goal. Let’s put this formula to work. Here’s what the shots look like in the hypothetical game when Matthews is on the ice:
|Shots on Goal||Missed Shots||Blocked Shots|
Then the numbers get added to the formula:
Matthews’ Corsi number = (2+6+4) – (1+7+2) = +2
With this Corsi number, you can then find Matthews’ Corsi ON statistic. This complicates the stat a little further by taking into account a player’s time on ice during the game. That formula looks like this:
Corsi ON = (Corsi Number)*(60 minutes)/(total ice time)
Let’s say Matthews was on ice for 17 minutes.
Matthews’ Corsi ON = (2)(60 minutes)/(17 minutes) = 7.06
Adding a time component to the stat helps explain the numbers further. The less time Auston Matthews is on the ice with a +2 Corsi number, the more impressive the number is.
The final component to the Corsi number is the relative Corsi. This is simply:
Corsi REL = (Matthews’ Corsi number) – (Leaf’s Corsi number)
So if the Leafs’ Corsi number for the game is +6 then it would look like this:
Corsi REL = (2)-(6)= -4
What do all of these numbers mean?
Basically, the Corsi number tries to give value to a player whether they are defensive or offensive. A defensemen won’t score the same amount of goals as an offensive player, but they can have just as good of a Corsi number.
Corsi numbers are not without their flaws. Just because a shot is taken does not mean it is a good shot, or ever had the ability to go in the net. It also does not judge the player on their own. Put Erik Karlsson on a more winning team and his Corsi number has the potential to skyrocket.
The Corsi number is not the end all be all for hockey statistics, but it’s the first step towards mathematically understanding the game.