THE MYTH OF SAVE %
Something I find very interesting is how sports writers and fans evaluate goaltenders, especially NHL goaltenders. When evaluating a goalie and figuring out whether he is having a strong stretch/season/career there are usually three stats that come into play: wins, goals against average, and save %. These seem fair as they imply that a goaltender who stops more shots will, in turn, win more games and therefor is more valuable to his team. But are these stats really as transparent as they seem? Let’s take a look.
Wins – I think it is pretty obvious that a goalie isn’t responsible enough for a team’s victories for wins to be a good evaluator of quality of goaltending. A goalie can let in 5 goals per game but if his team scores 6 goals per game, his team is going to win more often than not. On the flip side, a goalie can let in 1 goal per game but if his team never scores a single goal, that goalie is not going to win many games (just ask Cory Schneider). These are obviously extreme examples but they do show that a team’s ability to score (or lack thereof) does not translate to a goaltender’s actual performance. A goaltender’s performance can contribute to his team winning or losing but is not enough of an overall factor on a consistent basis for wins to be a good judgement.
Goals Against Average/Save % - This is where things get a bit tricky. Goals against average and save percentage are usually the factors we go to when it comes to deciding who is the best goalies in the game. The goalie who leads the league, in save percentage seems to automatically win the award (Rask, Price, Holtby, Bobrovsky, etc.) despite everyone and their mother knowing Carey Price is the best goalie in the world. There are a few minor issues with this and one major issue. The minor issue is that Cary Price is so much better than the next guy on the list that giving the Vezina to anyone else is like an MLB MVP award going to anyone not named Mike Trout. The major issue is that so little divides the top end goalies in terms of save percentage that the stat is mostly useless at this point.
Let’s look the save percentage of some of the best goalies in the world over the last five years*
* Minimum 45 games played. Stats for 2017-2018 taken on 3/26/18.
Just look at how much these numbers fluctuate from year to year. Let’s look at some of the top goalies in 2017-2018 in terms of save % and try to figure out some narratives for them over the past four years.
Pekka Rinne was widely considered to be one of the best, if not the best goalie in the world for a long stretch from 2008 on. He finished 4th in sv% in 2014-2015 then dropped all the way to 21st in 2015-2016. He must have been on the decline, right? Probably washed up and on his way out. Then last year he kicked it back into gear and finished 9th before tearing it up this season, sitting in 1st in save % at the time of this writing. So what caused that blip? Did he have an off year? Possibly.
Henrik Lundqvist is a similar example. He went from 7th in 2014-2015 and 2015-2016 all the way to 20th last year and now sits at 14 at the time of this writing. Remember the last two years when Lundqvist had “slumps” and everyone began to worry if the Kings’ time was up? Then we watched him in the playoffs and remembered that he’s still one of the top five goalies in the NHL?
And now my three personal favorites, Braden Holtby, Carey Price, and Cory Schneider – considered the three best goalies in the world by many. Price put up what can only be referred as video game numbers in 2014-2015 in his run to a much deserved Vezina and Hart Trophy. He was and is considered the best goalie in the world by pretty much anyone I know yet here he is this season with a measly .901 sv%, good for 22nd among qualifying goalies. Holtby was hot on his tail at 6th, 5th, and 2nd the past three years but dropped off to 21st this year with a .906 at the time of this writing. Cory Schneider finished 4th and 2nd in 2014-2015 and 2015-2016 respectively but dropped to 22nd in 2016-2017 and did not qualify due to missed time due to injury this season but sits at a measly .909 sv% in his time played this year. Are these guys, the most reliable and consistent goalies of the past five years, really having that bad of a season? The numbers would suggest so.
The fluctuations in “performance” does not seem to make sense on the surface but it only really doesn't make sense if you look only at the numbers. Those numbers can be effected by so many things that have nothing to do with a goaltenders' play. Last season, Sergei Bobrovsky had a .931 sv% - best in the league among goalies who played 45 games or more. Petr Mrazek had a .901 sv% - worst in the league for a goalie with the same qualifications. If we try to even out their playing time and assume each goalie played in 60 games and faced 30 shots per game, the difference in goals allowed over the whole season would be 54, or less than one goal per game. That means that if the two faced off against each other, every game would hypothetically be decided by one goal. And that is literally the best and worst goalie in terms of save %. Now consider how much better Columbus is defensively than the Red Wings and think about whether that difference might translate into one single additional goal per game. Even if it accounts for only one goal every two games, that would bump Mrazek into the middle of the pack or drop Bobrovsky from the Vezina to an average goalie (which everyone seemed to consider him in 2014-2015 because he had only a .918 save % - which translates to only 23 additional goals over 60 games in our made up scenario.)
On top of all that, there's the fact that it's a goaltender's job to control rebounds and control the game in front of him whenever possible. Guys like Lundqvist, Bobrovsky, and Bishop can put up good sv% numbers but they force themselves to make more saves each game because they tend to give up a lot of rebounds unlike guys like Price, Holtby, and Schneider who may not get as many shots against because they reduce second chance opportunities by controlling their rebounds better than anyone in the league. A goaltender could reduce their shots by 5 shots/game just by controlling rebounds on outside shots. Those 5 saves can really blow up their sv%.
If Bobrovsky had reduced the amount of shots against/game by 5 last season, his save % would drop to a .917, which would have tied him for 12th. This is also assuming that 5 shot reduction wouldn’t translate to less goals against, which it probably would.
This doesn’t even take into account the team in front of the goaltender. Think of this scenario, if Jonathan Quick only gets 20 shots per game, but those 20 shots are legitimate scoring chances, and he stops 18 of them. That 90% sv% isn't great but he may have made 3 or 4 saves in that game that no other goalie could make (i.e. this) and you can't say he had a bad game.
Meanwhile, another goalie, say Ben Bishop, may get 30 shots per game but they may be low quality shots from the perimeter and he may give up the rebound which leads to another shot and save he has to make. If Bishop lets in the same 2 goales and stops 28 shots, his sv% is now 93.33% which is insane. But say those two goals were his fault. Did he have a better game than Quick? Corey Schneider had an "off year" last year playing for a team who's best defensman is listed at 5'-11" and 34 years old. Do you think his 30 shots per game were the same quality of the 30 shots per game Corey Crawford was facing? Do you think Connor Hellebyuck's 30 shots against were the same quality as Devan Dubnyk's 29 shots against? Again, it only takes about 1 goal every other game to drop your sv% from best in the league to Cam Ward. That's one goal every 4 games to go from the top to the middle of the pack.
Also, consider this, if you had a proven guy like Quick or Price or Crawford, would you run the same exact defensive system as if you had a rookie/first time starter/unproven guy? Think of how conservative the Penguins got in their own zone when they had Murray behind them in the playoffs two seasons ago compared to how they were playing throughout the past two seasons when Murray began to prove himself.
I think too often, stats are used to create a narrative when the opposite should be true. I think the problem is that we should be looking at save percentage as the "what" instead of the "why". Save percentage tells you that a goalie "stopped more shots" than another goalie but it doesn't tell us why he stopped more shots.
The best comparison I can make for this is to think of NFL Quarterbacks. My favorite NFL team is the New York Giants and we have one of the most controversial QBs in the country. Half of Giants fans love him, half hate him, pretty much every fan of every other team thinks he sucks from what I can tell, but most analysts talk about him being in the elite level QBs. But his WRs are constantly changing and getting hurt, his O-line is constantly changing and getting hurt so he's always playing with different inexperienced guys in front of him. He's had great seasons statistically and terrible seasons statistically so which one can you trust? Do you just look at the stats and say "Wow, he was incredibly inconsistant" or do you look at the rest of the factors around him and say "hey, he's consistently great when he has a decent situation around him, it's the whole team that has been inconsistent"?
You can use long-term statistics to build an argument but what if a goalie is trapped in a terrible system or a terrible team for years? Or if a goalie is lucky enough to have a great system and defense in front of him for years? Sergei Bobrovsky put up terrible numbers in Philly then went to Columbus, won a Vezina and has had a number of solid seasons statistically since. But is he a great goalie that had a rough time on a bad Philly team or is he a bad goalie that is being lifted up by Columbus' defensive system? The numbers technically point to both completely contradictory narratives. There really is no way to know by analyzing the numbers because Bobrovsky is never going to experience the same conditions that he had in Philly again (knock on wood for him). If he goes to another team, those conditions will be completely different than what he had in Columbus. The only way to know is to watch him play.
I'm only 30 years old, not super old-school, and I'm not going to go all Trouble With The Curve on everyone and dismiss all fancy stats completely, but for goaltending more than any other position in hockey, numbers tell far from the whole truth. Too many factors outside the goalie's control go into those numbers and, realistically, there isn't even that great of a difference between the numbers of a great goalie and a subpar goalie. The eye test is even more important so that all these factors can be taken into account.