Winning and NHL Salaries: The Passion of the Cap

by Jose Mena

This is the first of a multi-part series examining sports economics under salary caps. This week features the NHL and its hard cap, what you should hope for your GM to do and why Flyers GM Paul Holmgren is taking a massive risk. There’s some math involved that we gloss over in this article; if you want the complete argument it’s all here. 

In the summer of 2009, the Pittsburgh Penguins signed Jordan Staal. With Evgeni Malkin and Sidney Crosby already on the roster, it brought their spending on centers alone to $23.5 million. Two years before, the New York Rangers signed Chris Drury and Scott Gomez to massive contracts, spending $21 million on centers, under a smaller cap. The Washington Capitals have committed to both Alexander Ovechkin and Alexander Semin, spending $20 million on left wingers alone last season. This year, the Philadelphia Flyers blew up their team in order to sign Ilya Bryzgalov for nine years at $51 million, making their investment in goaltending among the highest in the league. Are any of these approaches intrinsically better?

Under a hard salary cap, NHL teams have had to make difficult choices about where to allocate what money the cap gives them. The hard cap is unforgiving (ask the Chicago Blackhawks), and implicit in decisions to invest in a certain type of player (say, that big-game shutdown defenseman) is a decision to not invest in another type (a game-changing sniper on the wing). Under a system like this, teams have to find value wherever they can. To try to figure out where value comes from in the NHL player market, we compared each team’s yearly salary to the number of points they gained that year and looked for trends in the data.

Before we address the idea that spending on certain types of players can contribute differently to success, we need to examine the raw effect that spending has on winning in the NHL. In an uncapped system with highly unequal spending, the relationship between payroll and success is obvious – think about the Yankees and Red Sox in MLB, or all of the good European soccer teams you’ve heard of (roughly half of the Spanish economy is geared toward maintaining Real Madrid and Barcelona's solvency). In a system like the NHL or NFL where there’s a hard cap and the distribution of team salaries in much tighter, this relationship is less obvious. To determine the nature of this relationship, we found the correlation of each team’s total spending each year with their yearly number of points for the past four seasons. Note: all seasons represented by the year in which the Cup was awarded, so the 2007-08 season is “2008”. 

Correlations range from -1 to 1, with 1 representing a perfect correlation (teams that spend more always win more) and -1 representing a perfect anticorrelation (teams that spend more always lose more). That there’s no correlation whatsoever between spending and winning in the NHL in 2009-10 is fascinating; essentially, there was zero value to be had in spending an extra dollar. There is, however, one case in which we expect a correlation of zero: a league where all of the teams spend to the cap, because playing in the NHL is a zero-sum game. If team salaries were unusually tight in their distribution in ’09-’10, we’d expect a lower correlation than is normal.

This isn’t the case. The average spending (as a percentage of that year’s cap) stays stable from year to year, as does the standard deviation of that spending (the error bars on the graph above). What this means is that the Red Wings always spend to the cap and the Predators never do, and that these decisions are usually made on with a long-term plan in mind rather than any seasonal circumstance.

It’s here that we get lesson number one about successful teams. This is pretty obvious to anyone who’s followed the NHL, but if you want to win, you have to get production out of your entry-level contracts. The power of the rookie-level contract is such that it can make raw spending completely irrelevant to the success of your team (ask the Toronto Maple Leafs). The correlation between points earned and spending in ’09-’10 is practically zero because the restricted free agents (RFAs; Wikipedia has a good entry explaining RFA status) on entry level contracts before that summer are a bunch of excellent players. Here’s my all-RFA team for that summer using playoff performance, trophies won and the size of my man-crush on them as a metric:

Milan Lucic	Jonathan Toews	Patrick Kane
Loui Eriksson	Ryan Kesler	Bobby Ryan
Alexander Semin	Nicklas Backstrom	Devin Setoguchi
James Neal	Joe Pavelski	Steve Downie

Kris Letang	Nicklas Hjalmarsson
Duncan Keith	Marc Staal
Braydon Coburn	Erik Johnson

Cam Ward

Carey Price

Jaroslav Halak

Chances are you’re in love with these guys if they play for your team, and I’ve left out names like Jonathan Quick (tough decision), Antti Niemi, Cal Clutterbuck and Daniel Carcillo. This RFA team is insanely better than the collections of RFAs in any of the years in the sample; it’s a quirk of draft year and prospect progression that they all became RFA in the same year. Before the summer of 2010, you could have easily fit this roster under the cap and added somebody like Bill Guerin or, better, Scott Niedermayer to play the wily vet. You’d win the Cup. (Can you tell just how excited I am about the idea of this team playing together? Nicklas Backstrom on the third line!) On the other hand, not all of these guys were getting serious attention in ’09, so the gap between the money spent on them and their level of production isn’t as great. After the ’09-‘10 season, all of the players got market-value for their services, restoring the balance between their salary and their value. In addition, late-blooming Swedes like Henrik Sedin, Daniel Sedin and Henrik Zetterberg got significant raises that same summer over what they were making the previous year. So it’s not surprising that the correlation between winning and salaries was depressed in 2009-10 – most of these guys were making less than $1 million against the cap, and they’re all steals at a price like that.

Where else can you find value in the NHL? Two years ago, the bottom fell out of the NHL goaltender market. All of a sudden, goalies making $2 million per season were leading teams to the top of the league. Antii Niemi beat Michael Leighton to win a Stanley Cup, and J.S. Giguere and Tomas Vokoun didn’t make the playoffs. With this in mind, as well as anecdotes like Pittsburgh GM Ray Shero’s Crosby-Malkin-Staal approach from the open, we divided each team’s yearly payroll into how much they spent on each position on the ice (LW, RW, C, D, G) and performed a multivariate linear regression of those salary components against the points earned by each team. This is a fancy way of not-quite saying that we found the correlations of each positional salary with points, while making sure to weigh the relative importance of each positional salary. It’s like finding the trend-line to a scatter plot. We were able to use this to estimate the number of points gained per extra dollar spent in each position.

Two things stand out about this graph: the precipitous drop in the relative importance of goaltender spending after 2009 and the stability post-2008 of the relative importance of spending on wingers. Essentially, there hasn’t been large-scale undervaluation of wing players in the four years of the sample (look at my all-RFA team and tell me it’s not much stronger down the middle than on the wings). 2009-10 as a whole is something of an exception (as discussed above), and the defense market appears stable without it. The center market in 2007-08 was actually overpriced, with Mats Sundin ($7 million), Joe Sakic ($6 million), Saku Koivu ($4.75 million), Mike Comrie ($4 million), Sergei Fedorov ($4 million) and Robert Lang ($4 million) all either retiring or taking significant pay-cuts the following season to remain in the league. It’s difficult to make solid long-term predictions based on so little data, but in our sample it appears that there is less risk to be had investing in wingers and defensemen, and the center market was adversely affected by a glut of overvalued free agents.

What this means is that taking these extenuating circumstances into account, the effect of spending on any of the skater positions on success is more or less the same. This isn’t obvious. Most people think (myself included) that all their team is missing a certain type of player, be it a puck-moving defenseman, a playmaking center or a speedy agitator. That doesn’t seem to be the case – it seems to be much more important that the contracts your team’s GM hands out be intelligently valued, so you shouldn’t justify a bad splurge on a free agent on the grounds that “we always needed that kind of player”. I’m looking at you, James Wisniewski ($33 million over six years for this guy almost makes Mike Milbury look good).

It does appear, however, that it is ill-advised to chase the big-name free agent goaltender. There’s strong anecdotal evidence to support the idea that the NHL has shifted toward a paradigm where goaltending is de-emphasized in favor of stronger defending (San Jose, Detroit) or attacking (Washington, Chicago). For whatever reason, the goaltender market has carried with it substantial risk over the past couple of years, and that’s been reflected in successful teams investing carefully and wisely in their goaltending.

Maybe the 2011-12 Flyers will change that. Ilya Bryzgalov is certainly talented, and his play was underrated on an oft-forgotten Coyotes team. However, in the absence of a strong-RFA/weak-UFA explanation for the relative unimportance of NHL goaltenders to success, Paul Holmgren’s definitely swimming against the NHL tide. What’s less excusable is the Avalanche’s acquisition of and subsequent overpayment of Semyon Varlamov. Varly’s still young, but so far he’s shown himself to be inconsistent and last season was outplayed by two rookies. While it seems today like the Avs haven’t had more than a bag of pucks between the pipes since Roy retired (and for that matter, the Flyers have until now left the crease empty since Hextall left), it seems like this sort of overreaction spells severe trouble for them down the road.

On the other hand, Capitals GM George McPhee managed to put together one of the most impressive free agency periods in recent memory, in no small part due to his handling of Washington’s goaltending situation. Refusing to overcommit to a shaky young goaltender and then fleecing the Avalanche for his rights, and then making the steal of the year in signing Tomas Vokoun for just $1.5 million, McPhee immediately made himself a front-runner for GM of the year. Make no mistake: Vokoun has been an elite netminder who’s been criminally underrated while playing in forgotten markets in Nashville and Florida, and while age may be a concern, the Capitals are enormously improved by his experience in the crease. It’s this kind of shrewd investment that’s been characteristic of the new NHL, and what history we have is solidly on the Caps’ side.

(Thanks to NHLnumbers.com for compiling the data over the years.)