Despite their massive popularity among the analytics community, you would be hard pressed to find someone who believes that possession metrics tell you everything about the effectiveness of a player at even strength. Although they tend to heavily regress to the mean, some players seem to have the ability to sustain above or below average on ice percentages (or PDO) over extended periods of time. With that in mind, we can test the repeatability of on-ice shooting percentage and save percentage in order to determine how each metric regresses between two separate intervals of time. In this case we will look at the Sh%/Sv% of players who have played 1000+ minutes from 2007-08 to 2010-11 and compare that to their on-ice percentages from 2011-12 to 2013-14 (they must also play 1000+ minutes in this interval in order to be included in the analysis).
First, we will take a look at Forwards and their on-ice shooting percentages. Below is a graph that separates Forwards into buckets based on their on-ice shooting percentages from 08-11 (each bucket contains 30 forwards).
As expected, forwards with on-ice percentages that deviate far from the mean in 08-11 tend to move much closer to the mean in 12-14. However, it is important to note that forwards who have above or below average Sh%s in their first four seasons (08-11) still tend to remain above or below average over their next three seasons (12-14). This indicates that there is more to a forward’s on-ice shooting percentage, over extended periods, than just random variance. This is nothing new in the eyes of the analytics community. No one expects John Scott and Sidney Crosby to finish with equal on-ice Sh%s over time. Shooting talent and shot quality are both very prevalent in the NHL, the only issue is that it’s often difficult to differentiate between players who legitimately drive/anchor shooting percentage from those simply reaping the rewards or detriments of random noise (aka good or bad luck).
This next visual shows the same thing but instead for defensemen (each bucket still contains 30 players).
In this case, we still see regression but to a much greater extent. Regardless of what a defenseman’s on-ice Sh% looked like from 08-11, you should basically expect it to be average from 12-14. Also, it’s apparent that the on-ice Sh%s of defensemen tend to vary less so than that of forwards. Accompany that with the fact that Sv% regresses almost completely to the mean between a 4 and 3 year interval and it is safe to assume that defensemen have a minimal impact on their on-ice shooting percentage.
Once again we will look at forwards but this time their on-ice save percentages.
One quick glance and it’s obvious that forwards can’t control on-ice save percentage like they can shooting percentage. There may be a slight difference in 12-14 Sv% between the first bucket and the ninth, but remember that a good number of these forwards played primarily in front of the same goaltender between these two periods of time. Taking that into consideration, I’m inclined to say that forwards have virtually no control over their on-ice save percentage.
That leaves us with defensemen and their on-ice save percentages.
The results don’t seem to differ much from the forward data. On-ice save percentage regresses heavily to the mean yet again in spite of the fact that a portion of the sample played in front of the same goaltender. This outcome aligns with a multitude of other studies testing the repeatability of on-ice save percentage at the player level.
In summary, forwards have a measurable bearing on their on-ice Sh% whereas defensemen typically don’t. This is what separates Sidney Crosby from Patrice Bergeron and players like Patrick Kane from David Clarkson. Additionally, forwards and defensemen have virtually no control over on-ice Sv%. The fact that a defenseman’s PDO is practically beyond their control implies that their ability to drive/anchor possession stands as an optimal tool for evaluation (at even strength) on its own.
Usage-Adjusted Corsi% Leaders (Defensemen, 2007-08 to 2014-15, 400+ TOI)