We need to have a talk.
There’s been much consternation among Caps fans for a while now over the dwindling ice time given to Lars Eller and his nominal “third” line and the playing time afforded Jay Beagle and his fourth-line trio. Graphically, over time, it looks like this (via hockeyviz.com):
Over the first 45 or so games of the season, Eller had two large stretches in which he was getting more ice time than Beagle, and the two were pretty close in ice time during the other segment. Since then, however, Beagle has skated more than Eller. At five-on-five, that difference is pronounced - over the season’s first 44 games, Eller averaged 11.2 minutes at fives, Beagle 9.6; in the 29 games since, Eller’s ice time has dropped to 10.6, while Beagle’s has spiked to 10.7 (a not-insignificant 11% increase).
Presumably, part of that owes to the fact that third-line stud Andre Burakovsky missed 15 of those 29 games, and, in his place, the third line (generally with Zach Sanford and then Jakub Vrana in Burakovsky’s place opposite Brett Connolly) was simply not as effective as previous incarnations - Eller and Connolly played to a woeful 44.9 Corsi-For percentage over that span, per Puckalytics, and in the 42 percent range with both of those rookie forwards (but much better in limited time with Justin Williams... hmm).
During that stretch, Eller and Beagle each averaged 10.2 minutes at five-on-five. (For their part, Beagle’s line with Daniel Winnik and Tom Wilson played to a 50 percent CF in just under 100 minutes.)
So that makes sense - more ice time for a “more responsible” line that’s playing well, and a little less for a struggling trio featuring a rookie and a penalty-taking machine (seriously, Brett, cut that out).
But with the return of Burakovsky (impressive at that, thanks to Camp Nemo) and the playoffs looming, the Caps must be shifting back to an increased role for the third line and a reduced one for the fourth, right?
Not so fast.
Since Burakovsky’s return (a mere four games, so grain of salt and all that), Beagle has averaged 11.2 minutes at five-on-five and Eller 10.9. Rolling four lines with the playoffs just around the corner is a fine idea, and, in fairness, the fourth line has been crushing it - over those four games, they’ve posted a 64.3 CF percentage on the strength of impressive shot suppression. The third line over the same span? 64.8, largely fueled by high shot generation.
[This is where we pause to note that, extreme as this four-game sample is, if the bottom-six can consistently tilt the ice going forward, the Caps are going to be in terrific shape.]
But here’s where concerns over deployment start to arise. The Caps’ last three games were all at home, giving Barry Trotz the chance to match lines to his heart’s content. Power-versus-power? Go for it. Shutdown-versus-power? Your call, coach. And here are Eller’s and Beagle’s most-common opposing forwards in each of those games (via NaturalStatTrick):
- Eller: Matt Stajan, Cam Atkinson, Christian Fischer
- Beagle: Johnny Gaudreau, Nick Foligno, Anthony Duclair
In other words, Trotz has generally been playing Beagle and pals against better opposition. Gaudreau was second in five-on-five minutes for the Flames in that Calgary game, Stajan was 9th. Duclair was the Coyotes’ top forward in five-on-five minutes on Saturday night, Fischer was 10th. Interestingly, though, Atkinson skates on the Jackets’ top line, Foligno on the second. It’s also worth noting that over these three games, Beagle’s line has taken nine defensive-zone draws and 10 offensive-zone face offs, while Eller’s trio has a much more favorable 2/14 split.
And while it’s hard to argue with the recent results, it’s not necessarily a great idea going forward, and here’s why: Lars Eller is better than Jay Beagle, and Lars Eller’s line is better than Jay Beagle’s. Here’s a brief year-to-date comparison of the trios, again via the indispensable Puckalytics (click to enlarge):
Again, let’s note that a fourth line that can play to an even shot differential despite a 3:2 defensive-to-offensive zone start ratio and, of course, a goals-for percentage hovering around two-thirds is great. Really great.
But that line simply isn’t as good as the third line. How good is the trio of Eller, Connolly and Burakovsky? Per Corsica, among lines with more than 210 five-on-five minutes, the Eller line is third in score- and venue-adjusted Corsi-For percentage (61.6, trailing only Patrice Bergeron’s and Jeff Carter’s lines), fourth in Corsi-Against rate (behind two Bergeron lines (!) and a Henrik Zetterberg’s group), sixth in expected Goals-For percentage and first - let me repeat that... first - in actual Goals-For percentage, outscoring opponents 15-2.
Granted, the Eller line has had some favorable deployments, especially relative to Beagle’s line - just look at those zone starts (but let’s be careful not to assume that production results from relatively easy minutes rather than vice versa). But before you go citing score effects, let’s note that Eller has played more (in the same number of games) at five-a-side with the Caps up a goal than Beagle has (a stat that’s surprisingly led by... Evgeny Kuznetsov?). Overall, Eller has played roughly seven percent more five-on-five minutes with a lead than Beagle has (and even more when trailing, which is where some of that “score effects” argument is fair). Graphically, it looks like this (if you can make it out):
So Trotz is using Beagle in tough defensive deployments, generally, but not necessarily to protect leads. Interesting.
But, again, Lars Eller is better than Jay Beagle, and it’s not just Eller taking advantage of favorable minutes: Eller is better against just about everyone than Jay Beagle is. Here’s a look at each of them against different levels of competition at five-on-five (multi-year sample), Beagle first:
Quick primer on how to read this: green is good, purple is bad; the higher you look on the vertical axis, the tougher the (forward) competition; the further right you go on the horizontal axis, the better the (forward) teammates; and the bigger the circle, the larger the sample size (in minutes).
With that out of the way, what we see with Beagle is a guy who struggles quite a bit once competition gets above that 14 minutes per night mark - there’s not a lot of green north of that line, and a lot of purple, particularly in those bigger-minute bins. None of this is surprising - Jay Beagle is a fourth-line center. If he was treading water or having success against better players, he’d be, well, a better player. As it is, he’s well-slotted and can do a fine job in the appropriate role.
Now, for Eller:
(Note that the axes are slightly different in range.)
Eller’s chart looks a lot better, especially above that 14-minute mark we (somewhat randomly) used as a dividing point for Beagle. Dark green in that biggest bubble in the 14-15 range, even comfortably geen in that 16-18 range. Put as simply as possible, Eller has posted a (very) favorable shot differential against higher-quality competition, while Beagle is underwater against the same type of players. Add whatever context you want, but it’s hard to envision anything that bridges this chasm in any particularly meaningful way.
All of this is to say that the Caps’ have a terrific third-line center in Lars Eller, and a terrific fourth-line center in Jay Beagle. Those are great foundations for the type of bottom-six that helps teams win in the playoffs, and the Caps’ bottom-six certainly looks ready to play that part. But it’s likely going to be important that the third line remain the third line and the fourth the fourth - misunderstanding what each trio is capable of could be the difference between reaching some of those lofty goals the team has set for itself and yet another disappointing spring.