In Part Two of a series on passing stats, Ryan Stimson looks at how passing data relates to the more widely accepted shot-based data.
How passing statistics relate to shot-based data
If a player attempts more of his passes in the offensive zone, we can reasonably assume that he is driving play forward in ways Corsi does not capture; if a team completes several passes in the offensive zone, controls play, yet does not attempt a shot, Corsi tells us nothing from that sequence.
Whereas Corsi tells us a particular player’s or team’s shot attempts, passing statistics identify which players are “shot generators” that drive a team’s Corsi numbers. You may think that Corsi already includes this and it does, but not in the most detailed way.
When a player is on the ice, all of his individual shot attempts as well as the team’s shot attempts are recorded. We can identify the plus/minus in shot attempts when he is on and off the ice to see how much more or less of an impact that player has. However, if a player is on the ice for seven shot attempts, yet did not attempt a shot, how much impact did he really have?
The passing data allows to go a step further into assessing on-ice shot generation by looking at the total shot attempts that a player either attempted himself or set up for a teammate. When a player completes a pass to a teammate which results in a shot attempt, we refer to this stat as SAG, or Shot Attempt Generated. It can easily be thought of or renamed “Corsi Generated,” since that is essentially what it is.
With this stat, we can parse out who the true Corsi machines are by looking at the percentage of on-ice Corsi events that occur as a result of their direct passing stats.
Not only shot attempts, but shots as well
Now, while it may be useful to look at the numbers and identify just how much Corsi runs through a player via both their passing and individual shot attempts, we can take it another step further into determining the “quality” of these shot attempts. Whereas shot quality has been debated numerous times online, the idea of “shot attempt quality” is something that, possibly, can be measured.
When recording a completed pass followed by a shot attempt generated, should the puck miss the net or be blocked away, nothing further is recorded; however, should the shot attempt result in a goal or save by the goalie, another stat is recorded: Shot Generated, or SG.
I have been tracking passes for the New Jersey Devils this season and there is a difference between Travis Zajac generating a shot attempt versus Patrik Elias generating a shot attempt. The difference? 44% of the shot attempts Zajac generates result in a shot on a goal. 58% of Elias’ result in a shot on goal. So, on average, a shot attempt generated by Elias will reach the net 14% more often than one generated by Zajac. So apparently, Elias’ passing has a bit more quality to it than Zajac’s.
Why tracking by zone matters (again)
I now return to my emphasis on tracking passes by zone which I explained in the first article. I do the same for shot attempt and shot generation. This was not something I set out to do in October, but really came out of some discussion at In Lou We Trust with the readers. I noticed during games that passes originating from the defensive and neutral zones seemed to generate shots at a higher rate than passes in the offensive zone.
Think of the breakaways, odd man rushes, and a pass into space for a forward coming off the bench, what do they have in common? Generally, there is a lot of open ice for the recipient of the pass to work with. I determined that fewer bodies between shooter and goalie and the percentage that a shot attempt will force a goalie to make a save increase significantly.
By tracking from which zone these shots and shot attempts are generated from, we can identify which players are more adept at transitioning from defense to offense with efficiency. In the first article on passing statistics, I showed you what part of a single-game chart looks like. Here is the rest of it, which includes the shot-related data I track.
You will see the shot attempts generated by each player and position broken down by zone, as well as the percentage of the shot attempts that result in actual shots on goal. With this raw data, there is a tremendous amount of advanced analysis we can do. I will cover them in the future, but below are a few previews:
- What percentage of a player’s passes result in shot attempts?
- How frequently does a player generate opportunities for his teammates?
- Is a player more effective at generating opportunities from within or beyond the offensive zone?
- How much does a player truly contribute to his team’s Corsi? How much is a result of his shooting? Passing? Shot generation?
Ryan tracks zone exits and passing stats for In Lou We Trust, where he is a contributing writer. He is a long time fan of the New Jersey Devils.
Follow Ryan on Twitter at @RK_Stimp.