Introduction to Advanced Statistics in Hockey: Corsi, Fenwick and the Problems with +/-

Welcome to an article where I explore analytics in hockey. Now I know some of you are all sick of the analytic talk and think it’s garbage but it can be one of the most effective tool to calculate a player’s impact on ice. It’s a more advanced version of +/-.

I know you may not want to read this because you hate analytics but this might help you understand the offseason moves the team made. While some Panthers blogs try to explain this, they ultimately fail in writing an article explaining it. I did a lot of research on this so I know why we did it and it wasn’t to shut up people who point to our analytics and call us frauds and stuff.

+/- only calculates when a player is on the ice for a goal. Now a player can be awful and have a +/- of Positive 20 simply because he was at the right place at the right time. Meanwhile his Corsi and Fenwick could be terrible.


Corsi is calculated on a game by game basis. The formula is simple.

Corsi = (SF + MSF + BSF) – (SA + MSA + BSA)

SF= On Ice Shots For   MSF= On Ice Missed Shots For   BSF= On Ice Blocked Shots  SA= On Ice Shots Against  MSA= On ice Missed Shots Against  BSA: On Ice Blocked Shots Against

However, while regular old Corsi does have it’s faults  there is Corsi Relative which is basically the difference between an individual player’s Corsi and his team’s Corsi while he is on the bench. This is good for possession players.

Last year, when Ryan Lambert of Puck Daddy infamously published his Why the Florida Panthers are a fraud for now. He is basically pointing out the team didn’t have good Corsi statistics and are being outshot which means they would be relying heavily on goaltenders Roberto Luongo and Al Montoya. Now what most advanced statistics  don’t measure is scoring chances which can’t really be calculated but how they develop but those are far more complicated. It is unclear whether Lambert just published the article to hate on the Panthers or just try and make a name for himself by publishing a controversy or just stating the facts but in any case advanced stats don’t 100% determine a game’s outcome but they can be a useful tool in determining a game’s outcome.

Remember, good possession stats mean you have the puck a lot and that can wear down the other team to take penalties or just make bad mistakes that lead to goals being scored.

For example let’s say Aaron Ekblad is on the ice during the Panthers game against Colorado tonight. While he is on the ice, the Panthers get 8 Shots on Goal, Miss 5 Shots and they block 2 shots however the Avalanche get 5 shots on goal, Miss 5 Shots and block 3 shots while Ekblad is on the ice. Ekblad’s Corsi would be +15/-13 which overall is +2 for the game against the Avalanche.


This is by far the most simple advanced statistic to calculate. It takes out the blocked shots. Remember a problem with Corsi is that it counts blocked shots which like plus/minus can sometimes happen if you are in the right place at the right time which basically equals luck. Here is the formula for Fenwick:

(SF + MSF) – (SA + MSA)

For example, let’s say while Aleksander Barkov is on the ice tonight the Panthers get 13 shots on goal and miss 4 shots, while the Avalanche get 9 shots on goal and miss 5 shots. Barkov’s Fenwick would be +17/-14 which is +3


But you may ask, a player’s Corsi could be +9 one night against an opponent who struggles but the next night it could be -10 against a dominant team. Remember Corsi is calculated game by game. This doesn’t fully calculate the type of team but its the most useful tool.

Corsi Rel QoC is the average Corsi Rel (Player’s individual Corsi minus his team’s Corsi while he is on the bench) of opposing players determined by head to head ice time while Corsi Rel QoT is the average Corsi Rel of teammates weighted by the time they are on ice together. For example, the Huberdeau-Barkov-Jagr line usually spent most of their time in the offensive zone so they probably have a better Corsi Rel QoT.


As of October 22, 2016 at 5:07pm ET, the Panthers are the 4th best in the league at Corsi and 3rd best at Fenwick. Mark Pysyk has the best Advanced stats in the league while Jared McCann leads Forwards in the league.


One of the most hated trades of the offseason was Gudbranson for McCann. People complained the team was a bunch of weaklings to put it in an appropriate manner. Now, big body checks can force a player of the puck but at the same time, the player executing the check can’t take the puck right away which gives time for the player being checked’s teammate to come get the puck while a blocked shot is more likely to give possession to the team executing the blocked shot. Would you rather have a bone crushing hit in the corner but the opposing team takes the puck or a blocked shot a few feet in front of your goalie that forces the puck towards the boards out of harm’s way. While the physicality is entertaining, a blocked shot is a lot more effective in creating chances for your team.


Analytics can still be a risk. The Washington Capitals have had success with analytics being the best team in the league in both points and Corsi last year, but the Toronto Maple Leafs were also very high on the Corsi list but were the worst team in points last year. Analytics can either bring success or misery to South Florida, there is no in between. It’s like a big risk, big reward situation. The Panthers ownership decided to take the risk but it can be successful. I am quite confident in this ownership, because their business Virtu Financial is all math, all algorithms and have had great success and if they are successful they must know what they are doing.

In the next analytical article, I will cover zone and time on ice and stuff.


2 thoughts on “Introduction to Advanced Statistics in Hockey: Corsi, Fenwick and the Problems with +/-

Comments are closed.