Re-thinking What You Thought You Thought About Line-Matching

If EA Sports' series of hockey video games has taught us anything, it's that, given the chance, a rational coach would opt to play his checking line against the opponent's top trio of forwards. Sure, it might cost his own number one line minutes overall, but playing defensive-minded forwards against the opposition's big guns provides a measure of containment and allows virtual Bruce Boudreau to unleash the virtual fury of virtual Ovechkin-Backstom-Anyone on weaker competition when their shift comes up. In theory (and the theory, or a slight variation thereof, is well-conceptualized over at The Hockey Rodent), the result is a greater goal disparity per shift which, of course, means a greater chance at victory. Huzzah.

You'd think, then, that, in the interest of self-preservation (i.e. winning), all coaches would aggressively line-match in some similar fashion when given the opportunity. You'd be wrong.

One easy (but admittedly imperfect) way to figure out who's line-matching and who's just playing his guys regardless of whom the opposing coach is putting out there is to take a look at individual players' Quality of Competition for a given team: for the line-matching bench boss, there will be a wide spread here (as he tries to hide certain skaters by only playing them against weak opponents, play others against the best opposition, etc.), and for the line-roller, a much smaller differential (since he's not going to let the other guy dictate who he plays and when). 

So who's matching lines? For the 2008-09 regular season, the teams with the top five largest QualComp spreads among forwards with at least forty games played were Anaheim, Chicago, Calgary, San Jose and Pittsburgh (who might be even higher if they'd had Dan Bylsma behind the bench all season) and the bottom five were Boston, Phoenix, Columbus, the Rangers and the Islanders. But take enforcers out of the equation (since they're in a somewhat special situation when it comes to line-matching) and the top five are Chicago, Calgary, Anaheim, Ottawa and Pittsburgh, with Columbus, the New York teams, Washington and Boston bringing up the rear. In other words, Randy Carlyle and Joel Quenneville are big-time matchers, Boudreau (as we've previously noted) and Claude Julien are not.

That, in and of itself, doesn't tell us much - there were successful and unsuccessful teams at both ends of this spectrum and everywhere in between - these are just different coaching philosophies. But the inquisitive Caps fan and armchair bench boss is probably asking two questions right now: 1) Why doesn't Bruce Boudreau match more aggressively, and 2) Should he match more? We'll take these questions in turn after the jump.

As to the "why," you'll read a lot more about it in his book (you have ordered it, right?), but it suffices to say that when you have Alex Ovechkin at your disposal, you play him when you can - get too caught up in matching your checking line against the other team's top trio and you might look at the stat sheet post-game and realize that the League's best player was only on the ice for 17 minutes. Why let the opposing coach essentially dictate when you can utilize your best weapons?

Moreover, with few exceptions, over time top checkers are going to lose out to top scorers - match Boyd Gordon against Sidney Crosby and eventually you're going to be outscored. Skate Ovechkin against Crosby and Gordon against, say, Maxime Talbot's line and you've got a better chance to win or tie every shift (and win or tie every shift and you've gone a long way towards winning the game).

Further to that last point, matching a scoring line against a scoring line forces the opponent's top forwards to do something they may not love or be particularly adept at doing - backcheck and play defense. Whomever lines up opposite AO is going to have their hands full all night - why not give the other guy's stars something else to think about rather than letting them focus solely on how they're going to get around a few pesky grinders in the offensive zone? Of course, this cuts both ways - Ovechkin and his mates have to be committed to playing both ends of the ice when they're out against offensively gifted opponents. But it's always better to dictate the play than have it dictated to you.

Theoretically, the two sides of this argument - to match or not to match - could go back and forth all day. But to answer our second question - should Bruce Boudreau match lines more often - let's take a look at some data. Below are the cumulative plus/minus numbers (via HockeyAnalysis.com) for four Caps when matched at even strength against the top ten non-Cap scorers in the Eastern Conference from a season ago (note: data is only available for the top 75 individual opponents in terms of ice time against for each player, so stats do not represent each of the top ten non-Cap scorers for each of the Caps players below):

Now, obviously these are small samples, and a save here or a lucky bounce there and the numbers might look a little different. But behind these numbers are a few points worth noting:

• Since you're no doubt curious as to the combined numbers against the two Pens, Semin was +9.13, Ovechkin +4.82 (note: does not include Crosby data, due to a lack of head-to-head minutes), Steckel +1.75 and Gordon didn't have enough minutes against either to make his data available. In 39.43 minutes of combined even strength time against the Malkin and Crosby (which may include some overlap), Semin was on the ice for seven goals for and one against.
• Of the ten opponents we're looking at here, Gordon only had a positive plus/minus against Patrik Elias (one goal for, none against). Elias, Ray Whitney, Jeff Carter and Zach Parise were the only four of these opponents against whom Gordo played enough minutes to qualify. 
• Ovechkin's best plus/minuses came against Vincent Lecavalier and Martin St. Louis, Semin's against Malkin and Dany Heatley. Their worst differentials came against Chris Drury and Ryan Callahan for AO and Marty Reasoner and Whitney for Semin. This says, "play your best against their best," no?
• Steckel played within one goal of all opposing forwards with the exceptions of Scott Gomez, Markus Naslund and Reasoner (minus-two against each) and Brandon Dubinsky (minus-three in just 15.75 minutes). Gordon played within one goal of all opposing forwards with the exceptions of Carter and Scott Hartnell (minus-two against each).

So what's the big payoff? Well, for one, David Steckel seems (again, small samples) much more able to play against top competition than Boyd Gordon. And why do I single out Steckel and Gordon and not attribute a fair share of this to their respective linemates? Because each of them skated most often with Donald Brashear, Matt Bradley and each other. In other words, the linemates were virtually a wash (though Gordo did have a slightly lower QualTeam).

More to the point of this post, however, is the fact that Boudreau's reluctance to line-match is validated by the fact that his two top snipers have outplayed even the top competition they've faced - there's no need to sacrifice their minutes in hopes that a lower line can contain the opponent's top unit. Instead, fire at will... and Marc-Andre, and Tim, and Martin...