Beyond Big Four: Rethinking Vezina Trophy Numbers
Wins. Goals Against Average. Save Percentage. Shutouts.
This is the Big Four, history’s best attempt to quantify hockey’s hardest position. Just lately, it’s been receiving a lot more attention than usual because of a unique convergence:
You’re not seeing quadruple: Carey Price leads all four categories with only seven games left to play in the season. If he maintains these leads, he’ll become the first goaltender in the modern Vezina voting Era (1982-present) to do so.
Only Miikka Kiprusoff in 2006, Ed Belfour in 1991, and Patrick Roy in 1990 have ever come close, each finishing second in a single category. Each received the Vezina award as a consolation prize, and whether Price surpasses or merely equals them, he’ll claim the top goaltender honour as well.
Any NHL general manager from those seasons will tell you the same thing: those are the easy ones to vote for. But it isn’t usually so simple.
The Case of Recent Vezina Trophy Winner Mike Smith
In another reality, where the dominance of the Big Four had already been challenged, Arizona Coyotes netminder Mike Smith would have claimed that gleaming silver trophy.
It may be hard to imagine now, with Smith’s struggles receiving so much attention and his goalie coach set to find a new team, but he played an excellent 2011-2012 season, finishing tied for the league lead in save percentage. However, he trailed in goals-against average and overall record, not to mention reputation, and so finished out of the medals behind Henrik Lundqvist, Jonathan Quick, and Pekka Rinne.
Let me be the first to tell you this in 2015: Mike Smith was robbed, and you should be as retroactively outraged as I was when I dug into the numbers.
First, Smith’s “tie” in save percentage was a product of rounding: he actually edged out Lundqvist .9303 to .9298. A tiny difference? Sure, but an important one in such a tight race, especially considering the voting general managers historically place a lot of weight on save percentage rank. Smith also led in Goals Saved Above Average, the statistic that has successfully predicted seven of nine Vezina winners since the 2004-2005 lockout. Lundqvist finished second.
Nothing about the situation so far screams blatant injustice (aside from the fact that Smith wasn’t even nominated, of course). You could argue that Smith held narrow leads in important categories, but Lundqvist surpassed him in others, making either a solid choice. This all changes when you look at save percentage a little differently.
Though save percentage has (rightly) been recognized as the most important member of the Big Four, its inability to differentiate poor shots from great chances has been criticized recently. Wins, shutouts, and goals against average are all highly team-dependent measures, making them unreliable guides for judging goalies.
Sportsnet.ca’s Chris Boyle argues that save percentage as we usually calculate it is also subject to significant team effects, and it’s hard to dispute this. Poorer teams surrender more chances, and better quality ones, meaning that my .930 and your .930 might not reflect the same caliber of performance.
Two basic ways we now have for limiting team influence are 1) to include only even-strength shots, and 2) to adjust save percentage based on shot location. The first eliminates penalty-killing situations, where bad teams or teams with bad systems leave their goaltender highly exposed compared to others. The second takes into account whether shots come from low, medium, or high danger areas, and adjusts the save percentage accordingly.
I’m going to give you three guesses as to who lead the 2012 goalie class in even strength save percentage, adjusted save percentage, and adjusted even strength save percentage. If you didn’t guess Mike Smith, Mike Smith, and Mike Smith, you haven’t been paying attention. Lundqvist finished fourth, second, and eighth, respectively.
A Better Concrete Measure than GSAA?
Goals Saved Above Average is so useful because it combines save percentage and workload to produce a single, concrete figure: the number of goals your keeper stops compared to a league-average goalie facing the same number of shots. A refinement of this measure was introduced here (scroll to the end) earlier this month, showing Stephen Burtch’s improvement on the more basic approach. By replacing raw save percentage in all situations with adjusted save percentage at even strength, Burtch limited team effects as much as possible, giving us a more accurate picture of a given goaltender’s performance. I used his method to determine the adjusted even strength goals saved above median (or aesGSAM) for every season since the 2004-2005 lockout.
Returning to the 2012 season, you will, at this point, not be surprised to hear who finished first in aesGSAM.
Mike Smith led the league, while Lundqvist finished eighth. What may surprise you is the magnitude of Smith’s win. He finished with an impressive 31 additional goals saved. The second-place finisher, Brian Elliott, saved 15 additional goals. Lundquist saved 10. To put Smith’s distance from his rivals in perspective, as many goals separate him from second place as separate 2nd place from 37th. That kind of dominance has only been equaled in Tim Thomas’s brilliant 2011 season, where he easily took home the Vezina award.
Looking back at the last nine seasons, the Vezina-winning goaltender has finished first in aesGSAM only three times, and each time he also finished first in GSAA: Tim Thomas (2009, 20011) and Mikka Kiprusoff (2006).These were landslide victories where the winner topped most statistical categories. The other six seasons, however, must give us pause. If aesGSAM gives a more accurate picture with less team influence than other measures, shouldn’t it be the go-to metric for rapid evaluation? Have the best goaltenders, by the best numbers we have, been snubbed? Have some been unfairly undervalued, while others have been unjustly awarded?
The following table, showing the actual Vezina winner versus the aesGSAM leader may surprise you:
|Season||Vezina Winner (aesGSAM Rank)||aesGSAM Leader|
|2013-2014||Tuukka Rask (2)||Semyon Varlamov|
|2012-2013||Sergei Bobrovsky (3)||Jimmy Howard|
|2011-2012||Henrik Lundqvist (8)||Mike Smith|
|2010-2011||Tim Thomas (1)||Tim Thomas|
|2009-2010||Ryan Miller (3)||Tomas Vokoun|
|2008-2009||Tim Thomas (1)||Tim Thomas|
|2007-2008||Martin Brodeur (14)||Tim Thomas|
|2006-2007||Martin Brodeur (9)||Kari Lehtonen|
|2005-2006||Miikka Kiprusoff (1)||Miikka Kiprusoff|
While Tuuka Rask, Sergei Bobrovsky, and Ryan Miller all placed well in their Vezina years, Henrik Lundqvist, and especially Martin Brodeur weren’t even close. Had general managers had access to, and knowledge of these numbers, would it have changed the conversation, and ultimately, the Vezina winner?
While it’s too late for Mike Smith (though he has been receiving some praise for his recent anti-tank play), if the general managers take their voting seriously, they will have to consider the more advanced and accurate measures available, and put far less stock in the Big Four.
This season will be their fist chance to do so: who leads the aesGSAM race, and how could it affect the Vezina voting?
|1. Carey Price||27.7536|
|2. Cory Schneider||19.3905|
|3. Steve Mason||16.8756|
|4. Pekka Rinne||13.5857|
|5. Devan Dubnyk||12.6672|
Fine. You win again, Carey.
~ Paul Campbell teaches in the humanities program at Wilfred Laurier University, and also writes for thehockeychat.com. He’s a former CIS goaltender and women’s team goaltending coach at Mount Allison University. Originally from Cape Breton, Paul now lives in Guelph where he’s conducting experiments to determine which of his sons should don the pads (once they learn to stand on skates). Respectful feedback and spirited discussion are always welcome in the comments below.
[Statistics from war-on-ice.com and hockey-reference.com. Goalie leaders graphic from NHL.com.]