Numbers Guy: Goalies unpredictable, not easy to analyze
Minnesota Wild goalie Devan Dubnyk watches through his mask during a game against the Calgary Flames at the Scotiabank Saddledome on February 18, 2015. (Al Charest/Calgary Sun/QMI Agency)
To the naked eye, goalies seem to be the statistical world’s neglected species.
Over the past decade, as the hockey analytics community has continued to grow in quantity of members and quality of work, skaters have received the lion’s share of attention.
Nobody was taken aback when the league ignored goalies completely in its initial unveiling of “enhanced stats” on NHL.com last month.
Popular theories about why goalies have been overlooked include 1) the position’s numbers are highly unpredictable, and 2) the discrepancy between goalie stat lines tends to be slim.
“Goalies are so densely packed,” Rob Vollman, a hockey stats pioneer of sorts, told me this week. “The difference between the top goalie in the league and the bottom goalie in the league is so tight.”
While certainly not the only analyst diving into goalie metrics, Vollman has put his stamp on the matter at ESPN.com and elsewhere. The IT professional from Calgary wrote about things such as quality starts and home plate save percentage in the 2014 edition of his book series, Hockey Abstract.
Using even-strength SV% (ES SV%) as a starting point, let’s see what quality starts can tell us about this year’s goalie crop.
Of the 43 goalies that have spent 1,000 or more even-strength minutes in the crease, Carey Price stands alone. He’s the only netminder to break the .940 mark.
This is no revelation, of course, since the Vezina Trophy should already have the Canadien’s name engraved on its "2014-15" plate.
But, behind him is an 11-goalie logjam in the .939 to .930 range. Another dozen are between .929 and .919.
So, all tallied, that’s 24 goalies sporting similar save rates.
As you can see, Vollman isn’t taking a giant leap suggesting netminders’ numbers are “densely packed”.
It gets even muddier when you toss quality starts into the equation.
Much like baseball’s metric for evaluating pitchers, quality-start totals and percentages are calculated to isolate goalie performance from team performance.
A dozen NHL goalies have 30 or more quality starts this season.
Nashville’s Pekka Rinne paces the group with 41, banking a quality start (higher SV% than league average for the year) three out of every four games.
Journeyman Devan Dubnyk, whose season has been split between Arizona and Minnesota, has the same amount of quality starts (30) as Blues goalie Jaroslav Halak. It’s important underlying data since Dubnyk’s win-loss record (30-13) actually is a step down from Halak’s (34-16).
Hands up if, before the season, you had Dubnyk tagged as a fringe Vezina candidate.
“There’s always one Vezina finalist that everybody said was garbage,” Vollman said.
Home plate SV%, which uses shots taken only within an imaginary ‘home plate’ area in front of the net, looks at the influence of a shot’s location in relation to the goalie.
Shot quality is a topic of moderate contention among analysts right now. Some believe it plays a large role and make a convincing argument that not all shots are created equal.
Others feel the discrepancy between Shot A and Shot B isn’t significant enough to fuss about.
Either way, using a single stat to evaluate a goalie is probably not the way to go. With the gap so small, the more information the better.
“There’s this perception that there is a one-stat argument that, when you evaluate goaltenders, you just use that one stat: Even-strength save percentage,” Vollman said.
“But, in reality ... there are a lot of advanced stats that we can use now to study goalies.”
BIG TIME, BIG BUFF
Pending players' union approval, the NHL will adopt 3-on-3 overtime for the 2015-16 campaign.
Skilled players presumably will be the ones jumping over the boards in this unconfirmed scenario.
Considering 3-on-3 play is such a rare event in today’s game, we don’t have much data to work with. The next best thing is to look at who excels at 4-on-4 during tied games.
Since the start of the 2011-12 season, 114 players have played at least 60 minutes under these circumstances.
Eight of the 144 have produced a minimum of 3.0 points per 60 minutes. That smaller group is dominated by five highly efficient defencemen.
Hybrid blueliner Dustin Byfuglien, perhaps unsurprisingly, leads the charge with 5.0 P/60. With his all-world speed and shot, the ever-dynamic Alex Ovechkin (4.0) is next in line.
To round out the top eight, Duncan Keith (3.9) is joined by Anze Kopitar and Loui Eriksson (3.8), P.K. Subban and Alex Edler (3.3), as well as Kris Letang (3.0).
Although 3-on-3 will undoubtedly offer its own intricacies, differing from 4-on-4 in a few ways, it’s safe to say defencemen — offensively minded defencemen, more specifically — have the potential to star in next year’s new-look OT.
PADDING THE STATS
Are we not giving John Klingberg enough credit? The Star has 10 goals and 36 points in 53 games to pace all rookie blueliners. Klingberg, who skates 21:38 per game, is on Dallas’ top ES pairing and its top PP pairing right now ... According to ManGamesLost.com, the Blue Jackets have surrendered a league-high 444 games to injury this season. Far away at the other end of the spectrum are the Canadiens, who have been lucky enough to lose only 76 man-games ... This season, 25 of the NHL’s 30 teams have filled up their arenas to an average of at least 90% of their respective capacities. That’s an impressive ratio of arena-filling teams to those who don’t. The non-filler roll call: Devils (85.3% filled), Blue Jackets (85%), Coyotes (77.3%), Hurricanes (66%) and Panthers (64.6%).
Trending Up: Jordan Eberle, Oilers, F
Past 10 games: 4 G, 10 A, 14 P, 50.7 CF%
Trending Down: Joffrey Lupul, Maple Leafs, F
Past 10 games: 2 G, 1 A, 3 P, 46 CF%
*All stats current through Tues., March 18