April 24, 2011
(Ed. Note: Welcome to Stat Nerd Sunday, where we occasionally obsess over hockey numbers like a Dungeon Master obsessing over the level of his warrior elf. Here's Matt Barr, formerly of LCS: Guide To Hockey and Trolleytracks and now blogging hockey at Kertwang.me.)
But unquestionably, there have been some other stellar playoff performances in goal this year. Other shutouts, a lot of one-goal games, some in overtime or double overtime. How does one performance in goal really compare to another?
I wanted to try and answer that question by creating a Goalie Game Score -- a simple, one-number way to measure the quality of a performance in goal, that could be used to compare performances by different goalies in different games.
I wanted the Game Score to take into account what makes a game by a goalie impressive. It needed to include saves and goals against, venue, context (that is, score of the game) and whether the goalie faced a lot of power play time, and/or whether he had a lot of mental health time because his own team was on the power play a lot.
But I also wanted the relative weights of these factors to be simple. It's probably more impressive to shut out the Red Wings in Detroit than the Senators in Ottawa; but for clarity's and simplicity's sake, it's best to just add some fixed number of "points" for a start on the road.
Baseball has a Game Score for pitchers, created by Bill James, that has some of the same elements, and a similar emphasis on simplicity over complexity. It stirred some controversy in last year's MLB playoffs when the Game Score for Tim Lincecum's 2-hit, 14-strikeout win over the Braves in the Division Series scored higher (96) than Roy Halladay's no-hitter against the Reds a day earlier (94). (And Don Larsen's perfect game in the 1956 World Series, also 94).
I kind of dug the controversy because it got people talking about a relatively obscure statistic like Game Score, and because there should be an argument over whether Lincecum's or Halladay's performance was more dominant -- that is, it shouldn't just be an open-and-shut case that the no-hitter is better.
But let me explain briefly about the factors and weights I used to come up with this hockey version, and show you how the goalies are doing so far in the playoffs.
We award points for saves by taking the number of saves and dividing it by the number of minutes played. This helps contextualize long overtime games in the playoffs. Basically, it's saves per minute played.
We subtract 10 points for each goal against. We don't count empty net or shootout goals.
We add five points for a game on the road.
For each minute (or fraction thereof) of time the other team was on the power play, we add three bonus points. Scoring roughly (very roughly!) triples when one team has a man advantage, and it's unquestionably a more difficult and intense situation in which to play goal.
For each minute (or fraction) the goalie's own team had the man advantage, we subtract two bonus points. A goalie is obviously not under a lot of pressure when his own team is on the power play.
Finally, we give consideration to the final score of the game (again, ignoring empty net goals and shootouts). If the final score was tied (that is, a shootout game) or the goalie's team won by one, that's a 50-point bonus. A one goal loss is a 35-point bonus. A two-goal win is 20, and a two-goal loss is 10. The closer the score, the more important each save -- and each goal against.
It's possible to have a negative Game Score. 100 is very good.
Here are the top 25 regular-season Goalie Game Scores from 2010-11:
Craig Anderson's(notes) 47-save, 0-0 effort in Toronto by a good few lengths. It had everything we're looking for: hostile environment, lots of saves, no goals allowed, 12 minutes facing the opponent's power play and only six minutes of power play for the good guys.
Jonathan Quick's(notes) 40-save, 1-0 shutout in Philadelphia is quite a bit better than his 51-save, 5-0 shutout in Detroit on December 13 (not shown; Game Score: 102.26). Why? Mostly, the context -- we probably all agree that winning a one-goal game is harder, or certainly more intense, than a five-goal game.
Nowhere to be found among the top 25 -- or even top 50 or more? Ryan Miller. We meet him in the playoffs. Top 15 playoff performances to date, through games of April 22:
Because of the relationship between saves and minutes in this system, I'm looking at complete games only. This isn't just the Roberto Luongo(notes) Rule, it also excludes some very good relief performances from Boucher and Michael Leighton(notes), for instance. But as for complete games, here are the bottom five in the playoffs so far: