Capitals Spending in the Salary Cap Era

Last week, FiveThirtyEight published an interesting article on the strong relationship between spending and winning in Major League Baseball. Although baseball is where we got Moneyball from, most smaller-budget teams still can't consistently compete with top-half teams, Tampa Bay and Oakland aside. Spending still rules.

Prior to the 2004 lockout the NHL was also uncapped, but the introduction of the salary cap and floor changed the landscape. Non-player spending still mattered, but cap teams could no longer outspend each other on player cap hits—they had to simply be more efficient than the others.

At least, that's how it was supposed to work. The lockout presents an interesting opportunity to look at a league both with and without spending restrictions. Did the cap and floor truly level the playing field?

Using a similar approach as Davis and Lopez, we can look at this relationship in hockey. The values along the x-axis are standardized salaries—the number of standard deviations from league average. The values along the y-axis are wins. For 2000-04, ties were counted as half a win, and for the lockout-shortened 2013 season, the win total was prorated to be per 82 games.

Salary information via USA Today for 2000-14 and WAR-On-Ice for 2014-15 (though it's cap hits, not salaries, in the latter case).

Prior to the 2004 lockout, the best-fit lines were steeper—in three of the four seasons, steep enough that three standard deviations above the mean in spending might have implied 55-60 wins (assuming the relationship wouldn't have flattened out). After 2005, it's more of a mixed bag, although there still are seasons where the top spenders win that much more than the other teams.

Here's what it looks like if we divide the time period into three: 2000-04, 2005-10, and 2010-15.

(Click the image for a high-resolution version.)

We can see that the slope of the lines is a little less steep, and given the sample sizes involved, we can be fairly confident (think in the neighborhood of 80%) that that's the new reality—player salary spending now drives winning slightly less effectively as it used to. (The lockout season introduced some extra variation thanks to the abbreviated schedule, which is why it was separated.) The new relationship isn't as tight, either. Variation in salaries could account for 30% of variation in wins prior to 2004, but it's more like 20-25% now. Apparently, the floor and cap (in conjunction with rule changes designed to open the game up and the shootout) have helped a little.

Like in the baseball case, spending patterns and effectiveness have varied quite a bit between teams. Here's how the Capitals have stacked up over this time period.

The gray circles and line are league data over this time period. The Caps' line is steeper than the league line, meaning they've spent on players a little more efficiently than the league so far. (It would look even better if we looked solely at post-2004 data—the gray line would be less steep and the red line would be steeper, given the ineffective spending of 2001-02 and especially 2003-04.)

From this perspective, 2014-15 wasn't especially impressive by itself—it was a return to roughly league-average (and team-average) return-on-spending after a poor 2013-14.

Here's how other teams have fared.

(Click the image for a high-resolution version; you can also see each team individually here.)

There are several easy trends to spot at a glance. The Red Wings tend to be pretty good, regardless of how much they spend. Ottawa was good without spending much, but also pedestrian while spending more money. The Rangers haven't always spent efficiently (especially in unrestricted free agency), but have gotten much better as of late. The Blackhawks, Coyotes, and Hurricanes have tended to spend money smartly when they do spend.

Three intriguing teams here are the Predators, Sharks, and Maple Leafs. Nashville and San Jose's lines are similar in slope to the league average line, but are shifted up by about five wins—they don't get much more on the marginal dollar spent than the rest of the league, but they do start from a more efficient baseline. Toronto is Nashville's opposite—not so different from other teams on the marginal dollar, but seem a few wins below average to begin with. It seems safe to say the management shakeup of the last two years in Toronto and the relative stability of the front offices in Nashville and San Jose are all deserved.

Here's how Washington is shaping up for next season.

The Capitals will likely be a cap team yet again, and seem better set relative to the NHL than they have in a few years; if they can keep spending as efficiently as they have so far, they'll be competitive. But if they want to hit that next level, they'll need to spend a little more efficiently—whether that's through smart free agent signings, trades for undervalued players, getting big contributions from cost-controlled players, or in coaching and other non-player resources.