2013 NHL Stanley Cup Final - Game Six

“He’s grown up with our organization, and he’s been to the ultimate with our group,”
Chicago general manager Stan Bowman, on Corey Crawford

Back in the winter of 2003 or so, I can remember sports talk radio shows in Vancouver being almost exclusively about how the Vancouver Canucks needed not just a good goaltender to replace Dan Cloutier in net, but a “proven winner”. That last for most of the 2003-2004 season. Cloutier posted the worst even strength save percentage among starting goalies in 2001-2002 but is better known for allowing a half-court goal to Nik Lidstrom in Game 3 of the Canucks’ series against Detroit, and despite taking the first two games at the Joe, that goal seemed to turn the tide of the series, and Detroit wound up winning Games 3, 4, 5 and 6 to take the series.

Cloutier was slightly better in 2003, with a .917 EV SV%, 21st among starting goalies (although only 24 goalies qualified) but still he’s best known for his playoff performance—with a 3-1 series lead over the Minnesota Wild in the second round, the Canucks dropped the next two games 7-2 and 5-1, and then after taking a 2-1 lead into the third period of Game 7, eventually lost the game 4-2 and the series 4 games to 3.

Despite Cloutier’s just as miserable record in the regular season, all the radio callers wanted to talk about was Cloutier’s playoff performance throughout the 2003-2004 year. It didn’t matter Cloutier was having his one good year as an NHL starter, but he was basically a write-off for the playoffs. He’d played poorly in the postseason for the Canucks over a 22-game sample between 2001 and 2003, and was thus going to continue to be miserable.

Of course, that winter, there were only four active goaltenders with Stanley Cup rings, who fit the description as a “proven winner”: Ed Belfour, Dominik Hasek, Chris Osgood, and Martin Brodeur. Since 2003, those four goalies combined for a total of one Stanley Cup ring, via the only one of the four that probably won’t be in the Hockey Hall of Fame.

None of this really matters in retrospect. The Canucks didn’t get that “proven winner” but that was hardly the problem. Dan Cloutier was just a bad goaltender with just as shoddy of a regular season record as a playoff one. Goalies are judged for playoff performances now even if the playoff performance is consistent with the regular season one. Last week when I pointed out that Jonathan Quick had only a slightly higher EV SV% than Marc-Andre Fleury, the overwhelming response I got on Twitter was that Fleury sucked in the playoffs and Quick was good, which can mean we’ve finally reached the point where Fleury is no longer excused for his sins because he won a Stanley Cup once.

Is Jonathan Quick really good in the playoffs though? Dave Lozo has called Quick the best American goalie going right now, and Greg Wyshynski of Puck Daddy didn’t stop himself from calling Quick the best starting goalie in the NHL. Quick has been fine: over the last two seasons though, he’s just 9th in the league in EV SV%, and over the last three seasons he’s 16th among goalies with 3000 minutes played, tied with Carey Price and marginally better than Marc-Andre Fleury.

But Quick has two very good playoff performances to rest his heels on. After winning the Conn Smythe in 2012 with a .946 save percentage, he responded with a .934 in 2013 after struggling through injury for part of the season. 20 games during the Cup run was enough to define hockey’s view of Jonathan Quick, and when he responded with a very strong 2013 playoff run, it was enough to confirm our impressions: Quick is simply a good playoff goalie.

In his book Thinking Fast and Slow, Daniel Kahneman alludes to what he calls The Law of Small Numbers in a chapter about how humans naturally seek patterns in randomness. The simple explanation for Quick’s two-year playoff run is that Quick is good in the playoffs, but an even simpler explanation is that the two most recent years are a small sample:

Statistics produce many observations that appear to beg for causal explanations but do not lend themselves to such explanations. Many facts of the world are due to chance, including accidents of sampling. Causal explanations of chance events are inevitably wrong.

At the beginning of the chapter, Kahneman points out an example from Howard Wainer and Harris Zwerling, that the counties in the United States that have the lowest incidence of kidney cancer are rural and thus its “easy and tempting to infer that their low cancer rates are directly due to the clean living of the rural lifestyle—no air pollution, no water pollution, access to fresh food without additives.”

It’s just as easy to note that the counties with the highest incidence of cancer rates are also rural, which means “no access to good medical care, a high-fat diet, and too much alcohol, too much tobacco.” Both blips in the data lend themselves easily to the fallacy of narrative, but the explanations are inconsistent with one another. The correct statistical summary of why the highest and lowest kidney cancer rates come in small, rural counties is that because there is a smaller sample, there’s more of an opportunity for outliers.

It doesn’t matter, when looking to the future, what Quick has done to win or not win a Stanley Cup in the past. I looked at each of the last ten Stanley Cup-winning goalies (all different!) to find out what they did in the playoffs both before and after their Cup-winning seasons, statistically, pulled from Hockey Reference:

    SV% Before Cup Year SV% During Cup Year SV% Since Cup Year
2013 Corey Crawford 0.911 0.932 -
2012 Jonathan Quick 0.900 0.946 0.934
2011 Tim Thomas 0.926 0.940 0.923
2010 Antti Niemi - 0.910 0.909
2009 Marc-Andre Fleury 0.922 0.908 0.880
2008 Chris Osgood 0.910 0.930 0.926
2007 J.S. Giguere 0.932 0.922 0.902
2006 Cam Ward - 0.920 0.915
2004 Nikolai Khabibulin 0.916 0.933 0.898
2003 Martin Brodeur 0.918 0.934 0.914
Total 0.917 0.928 0.911

Here, we’re not dealing with samples of 600 shots against for a single playoff run, but rather over 6000 for each of the three “buckets” in the sample. Hockey is set up to have a winner every season, so naturally a goalie will get hot and win a Stanley Cup, but it appears to be independent of both their past and future performance. Future performance is ultimately how we have to judge hockey players. What they’ve accomplished in their careers up to that point is elementary. Quick is entering a ten-year deal, and it’s unlikely he outperforms that contract. Corey Crawford has seven years left on his term with the Chicago Blackhawks, and it’s unlikely he outperforms that contract. With the information we have about goaltenders, it is simply impossible to predict how they’ll do in certain situations.

“Has he won a Stanley Cup before, is he a proven winner?” is much less relevant to a manager’s job as “can he win us a Stanley Cup?” There are nine active goalies who were starters when their team won the Stanley Cup, and seven of them were first-time winners the years they won. The callers on radio call-in shows in Vancouver a decade ago seemed to have missed that point, that it doesn’t matter whether a goalie has won a Stanley Cup in his career if the goal is to win the next one contested.