One of the early concerning observations we've had regarding the Barry Trotz's Washington Capitals is how conservative the team has gotten with the lead. The possession numbers showed us early on that the Caps aren't making much of an attempt to create offense once they got a lead, and some recent blown leads late in the third period have brought those concerns back to the forefront.
Our pal Ed Frankovic has tackled the issue, suggesting that one of the keys to turning around the team's woes with the lead is to play Mike Green more when the team is ahead. Conceptually, it's simple enough - play your best players at the most important times. Ed and I had a brief Twitter exchange on the subject, and that led me to digging deeper on the extent to which Green has been benched while the Caps have had the lead and how it has impacted the Caps.
By way of backgound, I'm not a fan of rotating five defensemen unless the sixth defenseman is unavailable (injury, penalty, unexpectedly horrid play, etc.) because it either forces the coach to break up his pairs, or it forces the coach to sit a top-four defender every other shift . Take a look at the options for rotating five defenders (numbers indicate depth chart position; substitute names as you see fit):
Option 1 - Roll five defensemen: 1/2 - 3/4 - 5/1 - 2/3 - 4/5 - 1/2 ... Here, the coach is forced to break up his established defensive pairings and each defenseman plays with two partners, alternating partners each shift. This is really common in men's league where rosters are less reliable and chemistry is almost meaningless. In the NHL, however, defensive partner communication is crucial and if you are defending a lead this approach would introduce a greater risk of a too many men penalty or a miscommunication between partners.
Option 2 - Sub the fifth defender into an existing pair: 1/2 - 3/4 - 1/5 - 3/4 - 1/2 - 3/4 - 1/5 - 3/4 ... or 1/2 - 3/4 - 1/2 - 3/5 - 1/2 - 3/4 - 1/2 - 3/5 ... The coach ends up double-shifting one pair just as they normally would under a two pair scenario, but the other pair is broken up every other shift to accommodate the fifth defender. Again, there's an increased risk of a too many men penalty (my shift or yours?) and the increased ice time for the fifth defender is marginal. It essentially platoons the fifth defender with another top-four defender (presumably either the second or fourth, depending on whether the coach wants the first or third to partner with the fifth). More to the point here, if playing the fifth defender here was desirable, he wouldn't be the fifth defender.
Obviously, in certain situations the coach needs to make due with the available players. But when the full complement of defenders is available, it's probably best (and certainly easiest) to stick with the established pairs. That means if Green is going to play in the shortened-bench situations, he needs to out-perform a top-four guy. So, with that in mind, let's take a look at the data.
The first thing that stands out is that, per war-on-ice.com, Green actually plays a higher percentage of five-on-five minutes while ahead than he does while tied (didn't see that one coming, did you?):
That doesn't necessarily negate Ed's point - it's easy to make the case that Green should play more in all situations, including with the lead. Green is the best possession defenseman the team has, and keeping the puck away from the other team is obviously a good way to protect a lead. Green does benefit from some easier assignments than the other defenders, but he's been a great possession player for his entire career so there's no reason to think he's just living off of easy assignments.
By the way, while this data doesn't necessarily tell us that the Caps are trying to cut Green's ice time with the lead, GM Brian MacLellan said as much last week (at around the 16:45 mark), noting "if we're down a goal, Mike Green's playing a little more, if we're up a goal, he's not playing as much," and he's not wrong - per war-on-ice's data, Green is playing 31.5% of the minutes when the Caps are trailing by a single goal and 30.8% when leading by one - but the difference is probably less than you'd expect.
Still, it's strange to see that Green has a higher percentage of ice time with the lead than while tied; if the 28% of ice time he gets while tied is enough to make the Caps a very good possession team, why would the 30% he gets with the lead result in such drastically poorer results? Answer: it wouldn't. Here's Relative Corsi-For percentage for each defenseman, by score state:
The next thing that stands out is that Green's ice time with the lead has actually been increasing of late. Below are two charts showing this trend. The top chart is Green's percentage of ice time with the lead on a game-to-game basis; below that is the rolling 5-game average:
After a drastic dip in ice time while leading late in the 2014 calendar year, Trotz has gone back to Green much more since the New Year. Of course, lately the Caps have been giving up a troubling number of goals while leading, so let's hope that Trotz doesn't attribute that to Green's increased responsibility with the lead. The fact of the matter is any team that gives up playing offense when they get a lead is going to end up giving up goals, no matter how good the defense is. You simply cannot be as drastically below 50% possession as the Caps are while leading and expect to break even in the goals department. It seems unlikely that the recent blown leads is anything more than a bit of an unfortunate correlation with the rise in Green's time on ice while leading.
Of course, the Caps are in the business of winning hockey games, and "seems unlikely" isn't very comforting in the face of a frustrating losing streak. It would be nice to have some data to bolster the conclusion that Green's time on ice is not behind the rash of goals against late in games. Luckily, we've got some information we can look at.
Anecdotally, Green wasn't even on the ice for the third or fourth goals scored by the Nashville Predators (the first of which was on the power play) or Edmonton Oilers (the second of which was with the goalie pulled and can't be blamed on any of the defenders as it was simply an awful goal for Braden Holtby to allow). It'd be hard to pin the collapses on a player's defensive flaws when the player in question was on the bench during the defensive collapses. In fact, Green hasn't been on the ice for a goal-against allowed while the Caps had a third-period lead in the team's last 27 games, and only two all season.
More broadly, it looks like Mike Green might actually be achieving Barry Trotz's goals better than any other defenseman on the team. Despite Green's league-wide (and not totally undeserved) reputation that he's an offense-first guy with limited defensive ability, he continues to somehow manage to limit the opposition's offense (hint: the other team can't take shots if they don't have the puck). Below is a chart indicating the five-on-five Corsi-Against per 60 and Corsi pace per 60 (number of Corsi events per 60 of both teams combined) for all of the Caps defensemen this year when the team is leading (the story's the same whether leading by one or just leading, generally):
Green has the lowest CA/60 and CP/60 of any defender on the team. He's limiting opponents' shot attempts and slowing the game down (and, if Trotz really wants to limit offense, Green also has the lowest CF/60, which is not what we would have expected to see). Frankly, if you removed the names from the list I think most people would pick Green to be the most "Trotz-like" defender on the team given those numbers.
Visually, the results are similar:
Green appears to have a bit of an advantage in terms of zone starts, but the difference isn't as big as it looks; it's not enough to explain the substantial advantage Green has in CA/60 (and that zone start advantage sure isn't doing anything to help Jack Hillen out). Green has 13 more offensive-zone starts than defensive-zone starts in the 30 games he played in which the Caps had the lead at some point; that's one more OZ start every three games. Brooks Orpik and John Carlson are -5 and -6 in zone starts, respectively, each in 35 games. Every seven games they take one more DZ start than OZ start. Zone starts are not moving the needle here (if they ever do). Again, Green is limiting shot attempts against more than any other defender on the team, seemingly exactly what Trotz is looking for.
Playing Green more with the lead won't be a panacea (nothing is), but he's been the team's best possession defenseman for a long time, and simply put he makes the team better when he's on the ice. It may not always be pretty, and there's no shortage of media and fans ready to point out his every defensive gaffe, but at the end of the day he always seems to contribute more to winning than to losing. That's the mark of a good player, and Trotz shouldn't be so quick to hide Mike Green... even if working him into the late-game, lead-protecting rotation isn't always so easy.