Thursday, December 17, 2009

Fun with goalie numbers

In a long-distant post, I remember discussing goalie stats, and how the standards changed historically from the old, pre-expansion NHL, through the expansion and scoring boom of the 70's and 80's, and into the "dead-puck" era and beyond.  I can't find that post at all, sadly - it's not tagged.  But I do remember that in the comments, someone suggested that links to the pertinent numbers and goalies would have helped.

Well, this is that... kind of.  (I'm full of met expectations in this post, aren't I?)

The NHL's official stat pages for the keepers is a great resource for tinkering with numbers and comparing/contrasting.*  The guys I found and linked below, I've picked for particular reasons.  For example, you'll have heard that Martin Brodeur has won more regular-season games than any other goalie to play in the NH, and that he tied Terry Sawchuck's all-time-best mark for shutouts.  You may also know that after last night's game against Montreal, he is tied with the former Canadien, Patrick Roy, for most regular-season appearances by a goalie.  (If you click that first link above, you'll see it for yourself - at least if you click before the Devils play Ottawa on Friday, when Brodeur will pass Roy.)

*With one exception - they don't sort the career numbers based on any minimum number of games played, so the "all-time leaders" for the rate stats are pages of guys who played one game in net for their careers.  It can be a hassle to page through those and try to find the actual career guys.

Brodeur got his 103 clean sheets by allowing 2.21 goals per 60 minutes (the goals-against average is the goalie equivalent to a pitcher's ERA).  Sawchuck permitted 2.51.  Does that make Brodeur clearly better?  Is that more a function of the Devils' excellent system?  It's hard to say.  No complete records exist for the numbers of shots Sawchuck faced; we can't be sure what percentage of those he saved.  And from there, it's another step to compare his average to the league average.  Carl Yaztremski hit .301 in 1968, which sounds terrible to us, but he led the AL that year; it's possible that's more impressive than anyone who hit .335 in Coors Field during the steroid era.  To take Roy as an example: he led the league in GAA in 1991-92.  He saved 91.4% of shots against him, and surrendered 2.36 goals per full game; the league averaged 3.48 per team, per game.  Bob Sauve led the league twelve seasons earlier with an identical GAA (save percentage not recorded), with teams averaging 3.52 goals per game.  But nearly overnight, the game changed - in 1999-2000, Roy's 2.28 GAA (and identical .914 sv%) was only good for ninth, in a league where teams only scored 2.75 goals per game.  Was Roy actually better in 2000, or in 1992?

But the data get better every day, as do the tools used to understand them.  Tom Awad at Puck Prospectus has developed something called Goals Versus Threshold, for example.  In brief, it takes offense, defense, and goaltending, separates and quantifies each, and translates it into how many goals that contribution is worth to a team.  It's not easy to calculate, but it's a substantial move forward.  Folks also like the Corsi rating, named for its inventor, former goalie coach Jim Corsi, though like any stat, it's not perfect.**

** I have a book on goalkeeping which has contributions from Corsi, and it's excellent, much like the article on Corsi you can reach by clicking his name above.

As far as the basics?  Start with the unassuming, yet wonderfully-named, Doug Soetart.  Soetart finished his career with a winning record... but his GAA was 3.97.  It's the largest GAA I could find for any keeper who won at least 100 games, and more than he lost.  Even during the superball scoring of his time, it was rare for a guy with much worse numbers to keep playing.

You'll notice that in the few years Soetart's sv% was recorded, it was fairly unremarkable.  Ken Wregget, the longtime Flyer and Penguin (among others), was one of the first keepers who played an entire career with shots and saves officially recorded, and you'll see that he never reached .900 until 1995, and never got under 3 goals per game until three seasons after.  Not coincidentally, these were the beginning of the lower-scoring environments following the loss of half the '94-'95 season.  Wregget only recorded 9 shutouts for his career, and six of those were beyond this cutoff.

There's a lot of examples like this that suggest that a goalie is often dependent on the team in front of him to put up the sorts of statistics that catch attention; it's also part of why a number of people don't automatically call Brodeur the best goalie in history, just because he tops so many all-time lists.  Had the Islanders drafted him in 1990 instead of the Devils, for example, he wouldn't have played behind Scott Stevens, Scott Neidermeyer, Ken Daneyko, and Colin White, but behind Kenny Jonsson, Scott LaChance, Rich Pilon, and Radek Martinek.  Jonsson was a fine defender, but I don't think there's much of a comparison.  (Didn't pick those names at random, either - they're the four defenders who played the most games for those franchises since 1993, when Brodeur came up for good.  And fifth/sixth are Brian Rafalski and Tommy Albelin vs. Eric Cairns and Roman Hamrlik.  Oy.)

The most stark example I could find was old-time Charlie Gardiner, a Scotsman (aye!) who finished his career way back in 1934 with a 2.02 GAA... and a career losing record.  He was one of the last goalies to be named a captain (the NHL forbade it for many years, but recently permitted Vancouver's Roberto Luongo the honor); and won the Stanley Cup his last year while suffering from an illness that eventually killed him; he was named to the Hall of Fame in 1945.  That 2.02 is eye-popping, but the league's teams only scored 2.41 times per game in his last season.  He began his career before forward passing was permitted.  He also played for the so-called "goalless wonders," the Chicago Black Hawks.***

*** The 1906 Chicago White Sox were known as the Hitless Wonders... the '59 AL pennant winners weren't much better.  Maybe it's the city?

At least Gardiner won a title.  More modern (and less successful) examples include:

Doug Favell - 123-153-69, 3.17 GAA.  Started with the expansion Flyers, moved to the dreadful mid-70's Maple Leafs, and ended with the expansion Colorado Rockies.  Played from '67-'79.
Gary Smith - 173-261-74, 3.39 GAA.  A victim of one of the worst squads in NHL history, the late, lamented Oakland/California Seals.  Went 19-34-12 in '69-'70 despite a respectable 3.11 GAA; moved to Vancouver and went 32-24-9 with a 3.09 GAA and six shutouts in '74-'75.  Played from 1965-1980.
Tommy Salo - 210-225-73, 2.55 GAA.  Favored whipping boy of his own GM, "Mad" Mike Milbury (may his name be razed).  Won a gold medal in the 1994 Olympics, then moved to the Islanders, and eventually the Oilers; good enough to shut out 37 NHL opponents from 1994-2004.
Jamie Storr - 85-86-23, 2.54 GAA.  Seventh-overall pick in the 1993 entry draft, and tossed into action too soon behind some mediocre LA Kings squads.  Went 0-8 in his last NHL season to sneak under .500, and now plays in Germany.  Played exactly the same years as Salo.
Ron Tugnutt - 186-239-62, 3.05 GAA.  Another fine example of how the game changes: in 1990-91, Tugger had a 4.05 GAA in 56 games for Quebec; in 1998-99, in the midst of the dead puck era, he led the NHL with a 1.79 GAA in 43 games for Ottawa.  During that 1991 season, he faced 73 shots in a single game against Boston, stopping 70, for one of the greatest ties in league history.  Played from 1987-2004.

On the other end, there's Michel Larocque, who backed up Ken Dryden on the Montreal dynasty teams of the 70's.  He finished 160-89-45 with a 3.33 GAA from 1973-1984; but more accurately, he had two careers: Montreal ('73-'81: 144-48-31, 2.83 GAA,17 SO) and everywhere else ('81-'84: 16-41-14, 5.56 GAA, 0 SO).  Nobody could blame Larocque for playing second fiddle to Hall-of-Famer Dryden, but he failed to beat out Denis Herron in 1980; was supplanted by Richard Sevigny in 1981, and was moved to Toronto.  There he was outdone by backup Vincent Tremblay (out of the NHL by age 24!) and then lost his job to the underrated Mike Palmateer.  His last year was as the third goalie for St Louis, who had Mike Liut and Rick Heinz (who only played 49 games in his career, but now runs a goalie school).  At every stop Larocque was usually a half-goal or more worse per game than his fellow keepers.  The W-L-T record is wildly misleading when compared to his other stats and his teammates.

Two other notables are John Vanbiesbrouck, the all-time winningest US goalie, and Oilers/Sabres/Blues stalwart Grant Fuhr.  Both Hall-of-Famers from the 1981 draft class, with numbers which may pale next to modern keepers, but which were considerable when measured against their peers.  VBK played for mostly forgettable teams and won 30+ games only once, leading the league with 31 in 1985-86; Fuhr was the keeper for the Oilers dynasties in the mid-late 80's, and thus won more often; but his career GAA and sv% were worse and he had fewer shutouts (3.38, .887, and 25 SO against 2.98, .889, and 40 SO for Beezer).  He also had two fewer cool nicknames, though that wasn't his fault - "Grant Fuhr" is simple and memorable, and no reporter wants to spell "Vanbiesbrouck" five times, especially in the days when "cut and paste" involved an X-acto knife and Elmer's glue.

No comments: