We've talked about the idea in the last two posts about the concept of attempting to measure the total percentage of the time a team has possession in the offensive zone through the counting of all shots - shots on goal AND missed shots AND blocked shots. In other words Corsi.
But we have ignored a potential problem with corsi in doing so - why are we including missed and blocked shots? After all, aren't we penalizing teams that are good at blocking shots or forcing opponents to miss shots by including these shots in our possession statistic? In other words, if a team could block or force missed shots, it might not matter if it has possession less often because it minimizes the damage of that possession.
But how consistently can teams force missed or blocked shots? Let's start with missed shots: research into the ability of teams to consistently force teams to miss more of their shots has come up with an interesting finding: the amount of skill each team has at forcing missed shots is far smaller than the effects of random variance. To put that in layman's terms - while there may be a skill in forcing opponents' to miss shots, other factors are much larger causes of the amount of missed shots faced by a particular team. This actually makes sense if you think about it....missed shots are affected by things such as the shooters' skill, shooters' intentions, NHL scorers (think whether a shot-pass is considered a missed shot or not if it misses clean), etc. Again, in layman's terms, what this means is the following:
If a team through a certain amount of games - let's say half a season - has "forced" a large amount of opponents' shots to miss the net, it should NOT be expected for that trend to continue on for the rest of the season.
Now what does this mean? Well it means that removing missed shots from corsi would be a mistake - because missed shots forced is not really caused much by a team's own skill, removing them would be rewarding teams for things out of their control, and would ruin the predictive value of the stat. Remember, we want our possession metric - in this case Corsi - to show who is the best team (in certain ways) in talent, not who has had the best results.
The same is not true about Blocked shots:**
As you might imagine, there is a skill in forcing blocked shots.* This may seem obvious - but many obvious things in sports tend to not be true when looked into. More importantly, the skill of a team in blocking shots is a major factor in causing opponents shots' to be blocked. I realize I'm being a bit confusing here, so in layman's terms:
If a team has a particularly high blocked shot rate through a part of the season, we CAN expect this trend to continue and for that team to continue to block a lot of shots.
*Not to say there isn't a skill in forcing an opponent to miss shots, but that other factors are far more important in causing opponents to miss shots and these other factors are out of a team's control.
Enter Fenwick: Fenwick is similar to corsi - it's a measure of +/- of shots while a player or team is on the ice, but unlike corsi, Fenwick does NOT include blocked shots. Thus the formula for Fenwick looks like this:
Fenwick= (Shots on Goal For - Shots on Goal Against) + (Missed Shots For - Missed Shots Against)
Thus with fenwick, teams and players who are good shot blockers get rewarded for their efforts. These players thus have better fenwick ratings than corsi ratings. The same goes of course for teams - the Islanders for instance had a better fenwick (near even) than corsi because they were a good shot blocking team.
How useful is Fenwick? Well it's basically the most useful possession metric we have at telling which team is doing the best job at driving possession which in turn answers the question (mostly) of which teams are for real and which teams are likely to drop off. Yes, in essence, Fenwick is better than Corsi. So why not use Fenwick instead?
The answer is mainly sad: Fenwick is less available for use than corsi. Individual player fenwick isn't available at the best hockey stats site - behindthenet.ca - for reasons I do not understand. They are available at hockeyanalysis.com and timeonice, but both of those sites are more difficult to use. As a result, corsi is still the predominantly used metric, mainly due to availability.
**For a good piece of research on the influence of skill and random variation on missed and blocked shots, see this post by Sunny Mehta.
AND that's it for Fenwick. But before we move on from possession metrics, next time we'll talk about another statistic you've undoubtedly heard of before: Scoring Chances.
The Intro to Hockey Analytics/Advanced-Hockey-Statistics Primer so far:
Part 1: - What is the field of Hockey Analytics and Why Might You be Interested?
Part 2.1: - The Importance of Context Part 1 - Time on Ice
Part 2.2: - The Importance of Context Part 2 - Evaluating the Difficulty of Certain TOI through QUALCOMP and Zone-Starts
Part 2.3: - The Importance of Context Part 3 - Evaluating (and Compensating for) the Effect of Teammates via QUALTEAM and Relative Measures
Part 2.4: - The Importance of Context Part 4: The Concept of the Replacement Level Player
Part 3 - The Perils of Sample Size
Part 4.1 - Introduction to Hockey Analytics Part 4.1: Possession Metrics (Corsi/Fenwick)
Part 4.2 - Introduction to Hockey Analytics Part 4.2 - Possession Metrics: The Various Forms of Corsi Available on Hockey Sites