If you have visited any hockey related site or follow hockey on social media, you have probably noticed a recent increase in stats being bantered about that seem like they are written in another language. Advanced stats, metrics and analytics, or “fancy stats” as many have come to lovingly/mockingly refer to them, are really just the initial stages of a movement to bring hockey analysis into the modern era. The antiquated measures hockey has relied upon since the Dark Ages, such as Plus/Minus (+/-), are outdated and coming under heavy fire during this process. It’s not an attack on the history of the game, but instead a realization that there is a lot more to hockey than simply being on the ice when a goal is scored.
Personally, I think a better name is “expanded metrics” because most of the data being used has been available for a long time. The way data, such as shots and time on ice (TOI), are employed is where the growth or expansion comes in. So, while on their face the expanded metrics you are seeing more frequently may seem intimidating or too complicated for you to even try to learn about, they really are not. The purpose of this article and the related posts to follow is to pull the veil of obscurity off of expanded metrics so that more people who love the game of hockey can learn about and understand the game in even fuller fashion. You don’t have to become an expert in “fancy stats” to grasp their meaning. Many of these are as simple as batting average or on base percentage in baseball, but because they are relatively new, fans of all levels of knowledge just need a little education.
Many sports fans are familiar with football and at least some of the elementary stats used to keep track of a team’s performance. Time of possession is one of the more commonly used of these simple metrics. Time of possession allows fans to see how long a team’s offense controlled the ball during a game. Obviously, a team has a better chance of scoring if they control the ball. In football, the rigid structure of the game, with the offense on the field against the opponent’s defense, makes this stat easy to track. In hockey, this is not so easy to do. Hockey is more fluid with offensive and defensive players on the ice at the same time, but like football, a team is more likely to score when they have the puck (possession) than when they are stuck in the defensive zone.
Apart from teams of people watching the skaters with timers to determine how long each one has the puck, a way to track time of possession in hockey has been in need of a practical solution. Enter Corsi and Fenwick. These names may invoke an air of mystery but in actuality, they are just the names of the men who came up with the formulas to figure out time of possession in hockey.
Here’s a simple explanation of how each measurement works:
Corsi = shots on goal + missed shots + blocked shots
Fenwick = shots on goal + missed shots
Essentially, in order to shoot the puck, you have to possess the puck. Since we do not have the technology or another practical way to measure time of possession in hockey, we must employ a proxy. The measurement referred to as Corsi basically uses every shot toward the goal throughout the game. Every time some meatball in the nosebleeds yells “SHOOT!” and a player obliges, a Corsi event is recorded. Fenwick measures the same thing as Corsi but excludes blocked shots.
Expanding the stats we use to understand a player’s or team’s performance in the past, present and future opens doors to analyzing the game we love in all new ways. If we can identify trends in the numbers, we can predict with some degree of confidence what will happen in the future. We can determine the progress made by teams or players in different areas of the game. We can figure out if our team’s early season success is built to last for the duration of a long NHL season or just riding the coattails of some hot goal tending. The possibilities are as many and varied as the user wants them to be. Are these expanded metrics the first and last line of analyzing hockey? No. I doubt anyone would wholeheartedly propose that we can stop watching the games and just rely solely upon stats, but they do enhance our hockey experience.
Now that we have a grasp of what exactly Corsi and Fenwick are at their most basic level, we should explore how they are used. Both Corsi and Fenwick are counted as “For” or “Against”. “For” is a shot or event that happens while the player is on the ice that is on behalf of his team. “Against” is the same but for the opposing team. They can be applied team wide or by player. In general, Fenwick is usually regarded as a better indicator over a longer period of time. Corsi is a better indicator over a shorter period of time.
Example:
Patrick Kane is on the ice for 10 shots on behalf of his team during a game. The opposing team takes 3 shots while Kane is on the ice during the game.
Corsi For (CF) = 10
Corsi Against (CA) = 3
Kane would be a +7 Corsi (10 – 3 = 7) on the night.
Let’s say of the 10 shots Kane was on the ice for, 2 were blocked by players on the opposing team. The opposing team had 3 shots while Kane was on the ice but 1 was blocked. Because Fenwick excludes blocked shots from the formula, Kane’s numbers would look like this:
Fenwick For (FF) = 8
Fenwick Against (FA) = 2
Kane would be a +6 Fenwick (8 – 2 = 6) on the night.
To make this data easier to use, statisticians express a player or team’s numbers as a percentage. CF% (Corsi For Percentage) and FF% (Fenwick For Percentage) can then be easily compared among players, teams and games.
Example:
Jonathan Toews: CF CA CF% FF FA FF%
11/10/13 vs. Oilers 18 6 75.0 11 5 68.8
10/17/13 vs. Blues 9 4 69.2 7 2 77.8
11/6/13 vs. Jets 26 12 68.4 20 8 71.4
11/2/13 vs. Jets 9 14 39.1 8 8 50.0
(all stats via www.extraskater.com)
Remembering that a Corsi event is any shot toward the goal (SOG – Shots on Goal, Missed Shots and Blocked Shots), we see that Toews’ Corsi For (CF) varies greatly over these games as does his Corsi Against (CA). Focusing on these numbers alone could be misleading for the sake of comparing his performance with other players throughout the league, his teammates or even his own play from game to game. Using Corsi For Percentage (CF%) allows us to see how the numbers work together and remove the game to game variables that would otherwise be misleading or confusing.
Toews’ performance against the Blues (10/17/13) and Jets (11/2/13) were both a CF 9, but when the CA numbers are factored in and translated into a percentage, we see just how different those performances really were. Toews posted very good numbers (i.e. had a good possession game) at 69.2% against the Blues but had a disappointing 39.1% against the Jets.
When we remove the blocked shots from the equation, the Fenwick numbers take over. Toews’ best Corsi game of those listed above is third in Fenwick. Even the “bad game” in Corsi terms for Toews (vs. Jets 11/2/13) was a decent showing at 50.0% when viewed from the perspective of FF%. His CA of 14 becomes an FA of 8 due to the Blackhawks having blocked 6 shots while Toews was on the ice.
Using Corsi and Fenwick, particularly CF% and FF%, is just the beginning of the expanded metrics possibilities available to us. Many fans like to see these numbers just to confirm what they observed of a certain player or line during the game. Others dive deep into the numbers for more detailed analysis. However you choose to use them, if these expanded stats enhance your experience as a fan of the game, it sounds like a positive outcome to me.
It’s important to remember that we as hockey fans are going through a learning process because these expanded metrics are relatively new to the game. Don’t be afraid to ask questions for fear of “looking dumb”. We are all at different levels of the learning curve and also in uncharted territory so discussing these innovations and continually learning more about them is a shared experience that has the potential to make us all more educated fans.
In our next installment, we will learn more about these expanded metrics with an emphasis on Relative Percentages, Score Close and Score Effects.
