What Does Save Percentage Really Tell Us?

Editor's Note: Here's an excellent post by Van Ryn's Neurologist on save percentage.

As I'm sure we all remember, the Leafs played St. Louis on Thursday night. They had a 3-goal cushion early in the third, and then everything fell to pieces. Jonas Gustavsson looked, well, Toskala-esque. He had trouble seeing the puck. He had trouble finding the puck. Hell, he even had trouble touching the puck when it was sitting right in front of him. Clearly, he looked rattled. By the end of the night, he had posted a .886 SV%, slightly below his season average of .896, and well below what most would consider "good" goaltending. 

Then on Friday, James Reimer, during a 9-3 routing of the Atlanta Thrashers, looked poised (yeah, I said it). He looked calm. He looked sturdy. He looked like a quality NHL goaltender. And he had the.932 SV% to prove it, as well as 2 wins in his last 3 starts, and a very respectable 1.87 GAA to go with it. 

The large discrepancy in SV% had many painting Gustavsson (and Giguere) as a major problem in our first half struggles. Most of this argument hinges on his low SV%. If he would only stop more pucks, the theory goes, we would win more games.

Now, I'm not trying to advocate that Gustavsson should be given the start over Reimer in the next few weeks. I think Reimer has done an outstanding job so far and should continue to carry the workload until he proves he can't handle it. And Gustavsson likey is a little shaken lately, and probably needs some rest, coaching, or both. 

But, as I described in my last FanPost, save percentage has very little impact on number of wins. Other factors, such as 5-on-5 goals for/against and PK%, seem to play a much larger role in determining whether a team wins or not. (Scoring 5 PP goals sure doesn't hurt, either).

So why all the fuss about SV%? And surely SV% does mean something. But what? 

In doing a bit of research, I came across this great post by David Johnson over at HockeyAnalysis.com. Interestingly, this was posted just 2 days ago, so maybe he was also thinking about Gustavsson's save percentage.

I strongly encourage you to go over and read the entire thing. But there are a few things that I want to emphasize. 

First, take a look at this image:

What this image shows is SV% by score for each team over the last 3 seasons. The values are arranged from lowest to highest to make it easier to see the patterns. 

What patterns? Well what you should notice is that SV% is almost always highest when the score is tied (green dots). The obvious interpretation of this is that when the game is tied, teams play conservatively, which enables the goalie to make more saves.

More interestingly, I think, is that the SV% is much worse when teams are behind than when they are ahead. Again, as David points out, this suggests that when teams are behind they take more risks. These risks result in high-percentage chances on your own net (breakaways, etc.), which leads to a lower SV%.

Ultimately, what does this mean. Well, as David put it:

 It means we need to be careful when evaluating goalies (and probably shooters to some extent) based on save percentage (special team effects) or even 5v5 even strength save percentage because the game situations a goalie has been exposed to will influence the goalies save percentage.  A goalie on a weak team will have his save percentage lowered simply because his team is going to be trailing more often and be forced to take chances to create offense and thus he will be exposed to tougher shots where as a goalie on a good team who leads the game more than they trail a lot will not face as many tough shots. 

This is a really important fact. If the Leafs are trailing a lot, they will take more risks, resulting in high-percentage scoring chances against, and a bad SV%. Being ahead, too, can lead to taking the foot of the gas pedal (as we saw Wednesday with team Canada and Thursday with the Leafs). But being behind is probably far worse. 

But I think there's another more important fact to take away from this: SV% is not a purely individual statistic. If it were, SV% would be relatively constant regardless of the score. The fact that it varies so widely indicates that what the team is doing in front of a goalie affects his ability to stop the puck. In other words, some of the blame for a bad SV% needs to fall on the other 5 guys on the ice, and potentially, the guys behind the bench.

Case in point:

One interesting thing I noticed while doing all this was the Toronto Maple Leafs up by a single goal performance over the last 3 seasons.  While they were middle of the pack 5v5 game tied (16th in 3 year 5v5 game tied save percentage), they were downright horrific when they got up a goal.  They just couldn’t hold a lead.  The three worst single season save percentages when up a goal were the 2009-10 Leafs, 2008-09 Leafs, and the 2007-08 Leafs so they were three for three there.  Over the course of the past 3 seasons the Leafs posted an 88.4 save percentage when up a goal which was 3.44 standard deviations from the mean.  

So, for the last 3 years in a row, we have pretty average 5v5 save percentage. But when we got ahead, our save percentage was terrible. How terrible? Almost 3.5 standard deviations terrible! If a bad SV% were a mental illness, the Leafs would be certifiably crazy. But only while they were up a goal, not during 5 on 5. 

Why? What could cause such a dramatic drop in SV% when the Leafs are winning a game? I honestly have no answer to this question, but I have two wilde guesses. 

One possibility is that the entire team (including the goalie) chokes under the pressure of trying to secure a win. Between 2007 - 2010, Leafs fans were so desperate to get back into the playoffs, and I think the players must have known it. And lets face it, the ACC is a less than friendly place, even when up a goal. So, when faced with a lead, it's possible the team collectively craps the bed, with breakdowns in all 3 ends of the ice ending up with a final breakdown in net. 

Another possibility is coaching.  Hear me out. During the previous 3 seasons, the Leafs tried to play an aggressive fore-check style game, and refused to play a trap, or anything resembling it, when up by a goal. (I'm including Paul Maurice in this for the record, since this covers the 2007-2008 season as well.) Now I'm not advocating for a trap, but there has seemed to be a complete inability to adjust the style of play based on the score of the hockey game. So, while we continue to press and fore-check and send guys deep, the other team, coming from behind, are also taking a few extra risks. By failing to adjust for the offensive onslaught that comes with being up by a goal, the coaches may have inadvertently been facilitating scoring chances from the other team, resulting in a lowered SV% from our goalies. 

This is a huge assumption, I admit, and an unlikely one. One has to think that even the worst NHL coach would play for a win, and make a small adjustment to the style of play when up by a goal. In any case, for the previous three seasons, we have been horrendously bad when up by a goal. Again, some of this must be due to factors other than who's in net, since the 5-on-5 SV% was average.

But there's a silver lining.

The good news for Leaf fans is their 5v5 up a goal save percentage is much better this year: 95.6% (better than any team in any of the last 3 seasons), 97.2 for Gustavsson and 93.9% for Giguere so they are much better at maintaining the lead.  Unfortunately this season they can’t score well enough to get them a lead to protect.

Granted, this was before Thrusday's phenomenal collapse. But prior to Thrusday, Gustavsson had a 97.2  SV% when the Leafs were playing with a lead (5-on-5)! Some of this may be attributed to the fact that the Leafs so rarely had a lead that the sample size was so small. But, as David points out, one of the bigger problems was that the Leafs couldn't score prior to about a week ago. 

So what does this all mean. I think there are a few takeaway points:

1. SV% is affected by things other than who's in the net. When the Leafs are behind (which we have been a lot) the SV% is going to drop. The Leafs have fortunately played a few really good games in front of Reimer. It's going to be interesting to see how his SV% is affected when they play their first stinker in front of him. 
2. The Leafs need to learn how to get a lead AND play with it. The first trick is scoring early and often. It's been mentioned by others that we already need much more of that. But they need to continue to protect those leads and minimize the scoring chances against. So far this season, the Leafs are 13 - 0 - 2 when leading after the second period. So maybe some adjustments were made after last season. 
3. Gustavsson may be doing better than his SV% suggests. His .98 SV% when playing with a lead 5-on-5 suggests he has both the skill and the mental toughness to win games for us. He also has a very respectable .914 SV% at even-strength, even after Thursday's game. His biggest concern is the .802 SV% ok the PK. Clearly, he needs to make more of those key saves. But as I mentioned before, since SV% is affected by other factors, even that .804 SV% may be an unfair assessment of how he's played. Which brings me to:

4. In the words of David Johnson: We need to be careful when evaluating goalies based on save percentage. Keep that in mind.