Is Pekka Rinne Struggling Because of Fewer Shots?
Pekka Rinne is not having a great year.
It’s not the worst of his career. That distinction belongs to his injury-plagued 2013-14 campaign, when a post-surgical infection kept him out of action for months and he struggled to find his groove on returning. Rinne hasn’t been close to that level of bad this year, but a .923 at 5-on-5 is his second worst entry.
Naturally, especially with his selection to the All Star Game, there’s been a lot of analysis about Rinne’s numbers. Much of it is of a piece: Rinne faces a very easy workload, including a low rate of the worst kind of shots — shots from the High Danger zone (essentially the slot.) Because of that workload, his numbers have always been inflated and should be inflated. They’re not, which indicates just how bad this season has been for him.
If you ask Rinne, a decreased workload may be one of the culprits. As he told NHL.com, this season’s low shot counts have forced him to adapt in ways that have not been altogether comfortable:
“That’s been a huge adjustment mentally,” Rinne said. “You want to be a difference-maker, you want to do something to be able to help, and a lot of times this year after the game you feel like you didn’t do anything. It’s a terrible feeling.”
It can be even worse after giving up a goal.
“You want to get back in it right away, to have a positive effect on the game,” Rinne said, pausing for effect. “And instead you have to wait.”
Rinne wasn’t offering an excuse so much as an explanation, and is working to solve the problem.
“It’s so hard,” Rinne told InGoal Magazine. “You start to feel like, even though you try to stay focused, your mind sometimes starts wandering and things like that compared to when you know it’s going to be a busy night. I’m trying to take it five minutes at a time: to the next TV timeout, then regroup, reset, next tv timeout. [Using these ] five-minute segments and trying to break it down and set little goals for myself.”
He is also trying to handle the puck as much as possible (something Martin Brodeur used to do to stay focused amid low shot totals with the New Jersey Devils) to help him stay engaged in the game.
“I am trying to get every single rim, even when it’s up on the glass. I feel like that helps me, too. I try to stay vocal, try to communicate all the time, try to stay engaged in the game all the time.”
Rinne describes his style as “more flow and aggressive and movement and feeling the game.” Without that ability to feel the rhythm of the game, it’s more difficult to maintain focus and timing, which can lead to over-aggressive pursuit of the play and simply being in the wrong place at the wrong time.
The net effect of all of this is what appears to be a Rinne off balance, sometimes sitting back more, sometimes chasing more. The rhythm and flow of a classic Rinne game has been missing for much of the season, and the decisions he makes have been different. The drop in shots has had a real effect on how Rinne looks when he’s playing. It might just be having an effect on the results he sees.
That’s hard to prove, but the numbers are intriguing.
There’s very little correlation, for instance, between Rinne’s game by game save percentage and the rate at which he sees shots. The r^2 for shots against per 60 and save percentage by game for Rinne’s career is 0.089. That’s a statistical way of saying that a change in shot volume doesn’t always lead to a corresponding change in save percentage, at least not in single game increments. But if we zoom out a bit and group all of Rinne’s games by the rate at which he saw shots (of all types), a pattern does emerge:
It suggests workload may just be having a visible effect on Rinne’s results if we gather a large enough sample. In the long run, the fewer shots Rinne sees, the worse he tends to do.
It is important to note that this does not hold true for the NHL population as a whole.
If you look at all goalies together, you see that shot rates have almost no effect on save percentage. In other words, there is no pattern, not simply a pattern that is opposite to what Rinne’s experience suggests. What is true for one goalie may not be true for another.
In fact, there is even a question about whether we can “expect” a given save percentage to follow from High Danger (HD) workload. The link between the two isn’t nearly that direct.
Whether measured in terms of HD Shots Against per 60 minutes or in terms of the percentage of shots faced that come from the High Danger zone, there is no direct correlation between HD shot volume and unadjusted save percentage at the season level.
For the mathy types, I’ve drawn up some numbers [all data in this article is from war-on-ice.com.]
Taking all seasons since 2010-11 in which a goalie played at least 800 minutes at 5v5, linear regression returns the following correlations:
High Danger Shots Against per 60 to Unadjusted Save Percentage
p < 0.05
High Danger Shots Against as a percentage of all shots to Unadjusted Save Percentage
p < 0.05
What this says in essence is that there’s simply no predictable pattern for the effect of workload, whether HD or all shots, on save percentage. Sometimes goalies put up high numbers with harder workloads, sometimes lower.
It is clear that if Rinne performed at league average on all of the different kinds of shots he faced, his overall save percentage would be higher this season. There’s just no predictably causative relationship between the proportion or rate of High Danger shots an NHL goalie faces and how he will perform.
On the other hand, the results goalies get on those High Danger shots (HD Save Percentage) is extremely important to their overall save percentage. This correlation is one of the strongest in the hockey analytics world. High Danger shots are very important, but seeing fewer of them doesn’t by itself increase the likelihood that any given goaltender will perform better in any given season.
There has to be some mechanism that mediates between workload and results. How do we get from a measurement that has very unpredictable effects on results to one that has highly predictable effects?
There are a few possibilities. One is that there are other factors still unaccounted for that have a larger effect on performance than sheer workload. That is, there are multiple ways of experiencing a low High Danger workload or a low total workload. There are multiple ways that shots are dangerous. Those differences may be driving the results more than simple rate of these shots as a whole.
That detailed data isn’t collected by the NHL play by play files, so the only way to find it is to do manual tracking of every individual game. There are folks doing this, but the data is not yet complete.
A second explanation is that there are multiple ways that goalies interact with the shots they see. What is hard for one player may not be hard for another. In other words, how a goalie responds to his environment may be more important than the environment itself. Or perhaps strengths or weaknesses come to the forefront. To put it another way, individual difference in approach and technique may be the key.
We can see the goalie himself as a mediator between the environment and the results. A low shot count would be an input and technique would be the way that input is processed. Low event games and seasons, then, affect the goalie, whose technique may change, sometimes for the better and sometimes for the worse.
So something about Rinne’s individual experience this season is affecting him differently. Is it mental? Physical? Rinne says he feels healthy, but is age catching up to the demands he puts on his body?
Perhaps part of the answer lies in the difficulties of fitting his goaltending style with a low event environment. Rinne’s best seasons have come when he saw both more than 28 overall shots against per 60 minutes and more than 6.5 High Danger shots against per 60 minutes.
This year, he is at all time lows for both, and that may just be affecting him more than many realize.