wpostServer: http://css.washingtonpost.com/wpost

The Post Most: Sports

Complete hockey coverage
On Twitter Twitter: kcarrera and PostSports |  On Facebook Facebook |  E-mail alerts: Sports and Redskins |  RSS   RSS Feed
Posted at 10:53 AM ET, 01/12/2011

Statistical analysis: What are the Caps' best line combos?

We have all made the argument at one point or another along the lines of "Yeah, but he plays with better/worse linemates than [insert player here]."

One of the benefits of logging the scoring chances every game is that you get to see just how true this is. So when Capitals Coach Bruce Boudreau's lineups before the second Florida game had one goal scored -- combined -- during 5-on-5 play, I knew it would be a short-lived experiment. I got to thinking; what should the line combos be? I made the data available here so you can make your own picks. All data is during even-strength play.

The offense begins and ends with Alex Ovechkin, so it seems fitting to make sure that any top line has the Great 8 as an integral part.

Line Games TOI SCF/60 SCA/60 GF/60 GA/60
Backstrom-Knuble-Ovechkin 25 284.9 19.0 13.3 1.9 2.1
Backstrom-Ovechkin-Semin 24 163.0 21.4 14.7 4.4 2.6
Backstrom-Laich-Ovechkin 7 29.9 40.2 6.0 8.0 0.0
Backstrom-Fehr-Ovechkin 8 25.8 30.3 18.6 7.0 0.0
Johansson-Knuble-Ovechkin 6 17.2 17.5 24.5 0.0 0.0
Games = games where the combo appeared, TOI = time on ice, SCF/60 = scoring chances for per 60 minutes of ice time, SCA = scoring chances against, GF = goals for, GA = goals against

It's odd that the worst-performing line has seen the most ice time, presumably hoping that the power trio will somehow find its stride. It has been the Nicklas Backstrom-Eric Fehr-Ovechkin combo, however, that has produced some decent results, converting on 23 percent of its chances and not yielding a goal against during even strength. Could that be the best option for the top line?

If we go with Backstrom-Fehr-Ovechkin as the top line, that leaves Alexander Semin, Brooks Laich and perhaps Mike Knuble to anchor the second.

Line Games TOI SCF/60 SCA/60 GF/60 GA/60
Laich-Perreault-Semin 9 59.8 22.1 10.0 6.0 0.0
Johansson-Laich-Semin 6 41.2 17.5 10.2 0.0 1.5
Knuble-Laich-Perreault 4 21.0 22.9 8.6 5.7 0.0
Johansson-Knuble-Laich 4 15.5 23.2 11.6 0.0 0.0

With the trade of Tomas Fleischmann to Colorado, one of the young centers needs to step up --- at least until General Manager George McPhee makes another trade. The chemistry between Laich-Mathieu Perreault-Semin seems almost identical to that of Knuble-Laich-Perreault, the former benefiting from almost twice the ice time, so that combo gets the nod. The benefit here is that Knuble, with the recent injury to Semin, can be paired with Laich and Perreault almost without a loss in production. When Semin is healthy, I question where Knuble fits on this team. Locker-room leader? Sure. On-ice producer? Not this year.

The third-line and fourth-line options seem to give Boudreau the most flexibility, but it is an "all or nothing" proposition. Tough-to-find combos that have both produced and prevented goals and scoring chances.

Line Games TOI SCF/60 SCA/60 GF/60 GA/60
Bradley-Hendricks-Steckel 14 113.1 10.1 12.2 2.1 1.6
Beagle-Hendricks-Steckel 5 34.6 15.6 13.9 0.0 1.7
Bradley-Chimera-Johansson 7 31.6 11.4 24.7 1.9 1.9
Bradley-Chimera-Steckel 3 25.7 11.7 14.0 4.7 7.0
B. Gordon-Hendricks-Steckel 5 21.3 19.8 14.1 0.0 8.5
Bradley-Chimera-B. Gordon 6 19.6 21.4 12.2 3.1 9.2
Bradley-B. Gordon-Hendricks 3 18.3 16.4 9.9 3.3 0.0
Chimera-Johansson-Knuble 4 13.4 31.3 22.4 0.0 4.5
Hendricks-King-Steckel 2 13.4 4.5 4.5 0.0 0.0
Chimera-A. Gordon-Johansson 3 12.7 23.6 18.8 4.7 4.7

Matt Hendricks, Matt Bradley, Jason Chimera and Marcus Johansson along with AHL call-ups Jay Beagle and Andrew Gordon give Boudreau a blend of skill and grit, but Chimera is a cause for concern. The linemates Chimera has scored with leave his goals against per 60 minutes at a level that is simply not acceptable. Need scoring in a close game? Put Bradley-Boyd Gordon-Hendricks out there. Want to protect a lead? Bradley-Hendricks-Steckel is probably the right choice.

I can already hear the keyboards clacking with "Yeah, but could these sample sizes be any smaller?" to which I concede, yes - and no. It is true that some of the line combos have been on the ice longer than others, but when Boudreau reverts back to lines with which he is more familiar about 10 minutes into the second period, isn't that also judging on a small sample size? Or is that just how the game of hockey is managed?

Neil Greenberg also writes for Russian Machine Never Breaks. You can follow him on Twitter.

By Neil Greenberg  |  10:53 AM ET, 01/12/2011

Categories:  Statistical analysis

Read what others are saying

    © 2011 The Washington Post Company