Tuesday, June 25, 2013

The Caps' decline in puck possession

If you're reading this, you're probably aware that the Caps used to be an elite puck possession team but aren't any longer. For the uninitiated, Corsi is like plus-minus, but with shots instead of goals (and ignoring special teams play). And it's not just shots on goal—it includes misses and blocks, the theory being that while Corsi misses out on things like elite finishing ability, playmaking, and goaltending, it gives you a better picture of team quality than goals, on average. The Capitals used to be a top Corsi team. From Hockey Analysis (here and for the rest of this post, unless otherwise noted):
 
Season
WSH C%
League rank
C%, Tied
Rank
07-08
55.6
2
55.5
2
08-09
54.9
4
54.7
5
09-10
52.9
3
52.1
7
10-11
51.2
12
50.2
18
11-12
48.9
19
50.4
12
12-13
48.7
21
49.7
18

The final two columns are with the score tiedteams that lead a lot sit back and see a decrease in their Corsi%, and teams that trail a lot press forward and see an increase in Corsi%. 

I can think of a few explanations for the decline worth looking at right off the bat:
  • Schedule changesif the conference got better, it might explain part of the change. Remember that in 08-09, the NHL changed its schedule so every team would face every other team in a given season, sacrificing division games in the process (eight games against each division rival to six). It wouldn't be surprising if that was the cause of the 07-08 to 08-09 change.
  • Players getting worse as they age, especially Ovechkin.
  • Injurieslosing players like Nicklas Backstrom and Mike Green is going to hurt.
  • Coaching changesI can't think of a passive system that produces good Corsi, and as it is, players need time to learn the system to the point where it's instinct. (I don't have the knowledge of relevant players and systems right now to talk about this in anything more than broad terms.)
  • Roster changesnotably, losing possession monster Sergei Fedorov (and, more recently, Alexander Semin).
I haven't compiled the full play-by-play over this period, so I'm just going to use the tools from Hockey Analysis, Behind the Net, and Timeonice here to take a slightly macro look and see if there's anything odd that warrants a closer look.

First, we may as well look at the schedule. I think an easy way to do this is to take an average of opponents' Corsi%, weighted by number of games and unadjusted for location (home vs road).


Season
Opp C% (tied)
07-08
48.6
08-09
48.7
09-10
49.0
10-11
48.6
11-12
49.2
12-13
49.9

So schedule could be one factorI didn't subtract out the Caps' own games against these teams, so the increase in opponent Corsi could just be the result of the Caps' own decline—but if it is, it's a small one.

A big factor, on the other hand, is coaching. If George McPhee assembled this team for Bruce Boudreauand really, aside from Martin Erat, Jack Hillen, and Steve Oleksy, all these players were acquired while Boudreau was still aroundthen it only makes sense that this group of players is less than ideal for a different type of coach.

2011-12 tied C% under Boudreau: 55.4%
2011-12 tied C% under Hunter: 47.4% (although they finished the year strong)

Cognizant of the small samples we're dealing with, let's look at how players were affected individually by this change in 2011-12, via Timeonice:

Player
Boudreau
Hunter

C% tied
GP
C% tied
GP
Ovechkin
54.4
22
43.5
51
Backstrom
57.1
22
50.6
15
Semin
57.4
21
53.1
50
Brouwer
53.6
22
46.2
52
Knuble
56.6
21
40.9
45
Ward
56.4
20
49.8
41
Laich
56.5
20
47.9
54
Johansson
50.0
21
42.8
53
Chimera
60.4
20
51.3
51
Beagle
59.1
2
48.0
31
Perreault
56.3
12
55.2
47
Halpern
58.1
19
47.9
39
Hendricks
54.9
20
46.8
44
Eakin
54.8
11
44.2
18
Aucoin
0.0
0
52.0
24
Wideman
54.0
22
48.8
55
Alzner
56.3
22
43.6
52
Carlson
55.9
22
47.8
53
Green
55.4
8
52.0
20
Hamrlik
58.4
22
47.1
40
Schultz
50.0
21
45.6
32
Orlov
66.7
4
49.7
51
Erskine
46.9
8
48.9
17
 
The Bruce Boudreau part of the schedule was slightly stronger at about 49.6% opponent Corsi tied. (The play-by-play data is missing for three games in the Hunter part of the schedule.)

While context certainly plays a roleOvechkin isn't going to do well playing with Beagle and Brouwerit seems like the only players who might have avoided significant drop-offs are the guys who didn't play that much under one of the two coaches, i.e. the guys who couldn't have a drop-off in the first place.

I can't say for sure that injuries made a huge difference in 2010-11:

Caps w/Green C% (tied): 50.6%
Caps w/o Green C% (tied): 50.4%

And in case you were wondering, from TOI (keeping in mind the small samples we're dealing with here):

2011-12
Boudreau
Hunter
Player
GP
C% (Tied)
GP
C% (Tied)
w/Backstrom
22
55.4
15
48.6
w/o Backstrom
0
N/A
42
47.0
w/Green
8
59.6
20
49.3
w/o Green
14
53.2
37
45.9

(Sidenote: what exactly was Hunter trying to accomplish? I wish he hadn't broken what Boudreau had built, especially if he knew he would not return. He made the team much worse, although it was probably headed in the wrong direction anyway. Remember, as much credit as Hunter is given for taking the Caps to within a game of the conference finals, he was also one shot away from losing by one goal in Game 7, only as good as Boudreau's worst.)

I suspect this is more complex than just coaching, though. After all, the players are pros and have played for many a coach. I doubt a few months playing one system would destroy the chemistry they had...

...although the effect is so pervasive that it's hard not to blame most of this on coaching, which seems like it suppressed what these players spent so much effort developing and executing at a high level.

Looking at Ovechkin, specifically:


Season
Ovechkin C%
07-08
58.0
08-09
57.4
09-10
57.9
10-11
53.9
11-12
47.5
12-13
49.2

He's been getting worse, to no one's surprise. It hasn't just been him:


Season
C% w/o AO
07-08*
54.3
08-09
53.5
09-10
50.6
10-11
49.7
11-12
49.6
12-13
48.6

*Excludes games with Huet, since his Hockey Analysis numbers include his time with Montreal. This number could be a bit higher. Timeonice says it's 54.3%

And going by lines:


Season
Top line
2nd line
3rd line
4th line
07-08
58.0 (Ovechkin)
56.2 (Fleischmann)
52.1 (Laich)
54.8 (Bradley)
08-09
57.4 (AO)
57.6 (Semin, no AO)
55.6 (Nylander, no AS)
47.3 (Gordon)
09-10
57.9 (AO)
50.5 (AS, no AO)

53.1 (Steckel)
10-11
53.9 (AO)
52.9 (AS, no AO)
49.0 (Chimera)
50.1 (Hendricks)
11-12
47.5 (AO)
53.1 (AS, no AO)
50.6 (Ward, no MH)
47.7 (MH)
13
49.2 (AO)
44.9 (Ribeiro, no AO)
50.0 (JC, no MR)
46.4 (Beagle)

(I tried finding a simple proxy for each line, a combination of plenty of games played and playing consistently on that line. I couldn't think of one for the 3rd line in 09-10, since I felt the forwards outside of Ovechkin, Backstrom, Knuble, and Semin all played throughout the lineup regularly. I'm hoping that context doesn't play a huge role here, although it's probably worth noting that Boudreau used his fourth line in a checking role with harder zone starts, and gave Ovechkin and Semin pretty nice, but not Sedin-esque, zone starts, at least until 09-10.)

Although the second line suffered after losing Fedorov in 2009, it still put up respectable puck possession numbers until this year (when Mike Ribeiro took over anchoring the second line). But in 08-09, the Capitals had three great Corsi lines; in 2011, merely two good ones and two average ones; in 2012, one good, one average, and three below average; in 2013, two average and two bad ones. This is bigger than merely losing a major player or two—all the lines suffered.

Here it is going player-by-player (min four seasons, minimum 300 mins TOI to qualify as a "season"):

Player
07-08
08-09
09-10
10-11
11-12
12-13
Ovechkin
58.0
57.4
57.9
53.9
47.5
49.2
Backstrom
59.4
56.2
56.7
54.2
51.0
51.3
Semin
55.9
57.4
54.1
54.1
51.9
(CAR)
Laich
52.1
54.4
50.8
52.3
47.1
(inj)
Green
57.7
57.2
54.4
52.6
52.3
48.6
Carlson
(USHL)
(OHL)
54.0
52.4
47.7
49.2
Alzner
(AHL)
56.1
(AHL)
51.9
46.5
49.0
Erskine
51.9
52.6
51.0
49.8
48.1
45.5
Schultz
54.4
52.0
53.9
49.1
47.9
46.9
Poti
56.5
52.8
53.9
49.5
(inj)
(inj)

The orange is a drop off of 1.5 percentage points or more, green an increase of a similar amount. Given that the declines are across the board, I'd say that's a function of system changes/effectiveness and/or a decline in the quality of the blueline as a transition unitskating (Shaone Morrisonn wasn't very good, but he could skate with the puck) and puck-moving (remember that when Mike Green wasn't on the ice in 08-09, it was likely that either Tom Poti and Brian Pothier was).

For a comparison, take a look at Detroit and San Jose, the only two positive puck possession teams over this entire six-season period:


Player
07-08
08-09
09-10
10-11
11-12
12-13
Datsyuk
63.3
60.2
57.4
55.8
59.7
57.3
Zetterberg
63.9
58.7
56.3
53.9
56.6
54.7
Holmstrom
64.2
59.2
58.8
56.0
51.7
(ret)
Filppula
57.5
57.3
52.5
50.2
52.1
55.5
Franzen
59.8
60.9
59.9
52.8
57.6
53.2
Cleary
58.7
57.0
55.5
52.2
55.2
53.6
Hudler
60.6
58.7
(KHL)
52.8
54.6
(CGY)
Lidstrom
59.9
58.5
56.9
51.1
56.9
(ret)
Kronwall
58.6
58.7
54.8
54.2
53.2
53.7


Player
07-08
08-09
09-10
10-11
11-12
12-13
Thornton
58.2
55.5
51.4
53.6
55.6
54.4
Marleau
53.7
55.5
53.2
52.9
53.7
52.9
Couture
(OHL)
(OHL)
(AHL)
56.1
54.5
52.7
Pavelski
52.8
55.5
53.2
56.7
54.8
52.9
Clowe
56.5
56.4
52.5
57.1
52.0
(NYR)
Boyle
(T.B)
56.4
54.1
52.8
51.8
55.2
Vlasic
52.0
53.5
50.8
53.8
52.7
53.7
Murray
51.5
53.4
51.3
52.3
46.7
(PIT)

Detroit was insanely good in 07-08 and 08-09 (aside from in goal) and it's not surprising that the Red Wings couldn't maintain that stellar level of play. What I do find curious is all the green boxes for the 11-12 season. While the Caps do have assorted green boxes spread around, the Red Wings got them all at once. Both Datsyuk and Zetterberg got sizeable zone-start bumps, and the rest of the guys for Detroit that I listed play a lot with those two. So their success must have come partly at the expense of the bottom of the lineup for Detroit. The fact that they were able to rebound despite their age is encouraging for the Capitalsthey can rebound, tooand, I think, a sign of good coaching. The bad news for the Red Wings is that that'll probably not be enough to counter the effects of age for much longer.

San Jose, meanwhile, like Washington, had a coaching change that led to utilization changes that looked like they were for the worst, initially. Todd McLellan took over from Ron Wilson for the 08-09 season, and in 09-10 decided to start using Joe Thornton and the rest of this top players in a power-versus-power role, even though the Sharks were the second-best puck possession team in the West in the previous two seasons (behind Detroit, obviously). While Thornton initially had difficulty, after a season or two he got the hang of it. (I hope Adam Oates can work the same magic with Alex Ovechkin and Nicklas Backstrom.) This year, Logan Couture replaced Thornton in that role, and Justin Braun replaced Dan Boyle alongside Marc-Edouard Vlasic on the Sharks' shutdown pair. (Meanwhile, Douglas Murray stopped playing with Boyle and Thornton in a shutdown role, getting replaced by Vlasic, and lost a step he couldn't afford to lose.) San Jose is not reliant on old players as Detroit, but the Thornton-Marleau-Boyle trio is still pretty old, and they were able to rebound, too. It's possible for the Capitals—Adam Oates just has to figure out how. (Hint: Sedin starts would help.)

The Sharks have had the versatility to have players swap roles if one becomes better suited for it, and the two-way talent to beat the opponent's easy minutes and to hang with the opponent's best players. Backstrom fits that mold, but aside from him the Capitals don't have those sorts of players.

Also worth noting from these two teams: they kept around the players that made them terrific possession teams in the first place, so even with age they were still great, if not top, possession teams. (The Capitals dropped at least two and dropped further, Corsi-wise.)

Pittsburgh, strong since 08-09, looks more intimidating:

Player
07-08
08-09
09-10
10-11
11-12
12-13
Crosby
50.4
49.6
54.7
53.2
57.7
56.7
Malkin
47.9
46.2
51.1
56.7
56.6
56.5
Dupuis
(ATL)
50.3
49.5
50.0
53.4
52.4
Kunitz
(ANA)
(ANA)
56.2
53.9
57.8
54.8
Staal
46.7
49.8
54.2
53.1
54.5
(CAR)
Kennedy
47.2
53.2
54.8
55.6
57.6
49.3
Cooke
(WSH/VAN)
48.0
53.3
52.1
52.7
44.1
Letang
46.0
50.8
55.2
55.2
54.3
54.0
Orpik
45.1
45.8
52.1
50.7
52.3
44.6
Goligoski
50.8
52.6
52.4
(DAL)
(DAL)
(DAL)
Martin
(N.J)
(N.J)
(N.J)
54.6
55.2
50.1

Nothing surprising here. Dan Bylsma improved the team a lot when he took over in 2009. Kris Letang and Jordan Staal were improving rapidly even before that, and it looks like Evgeni Malkin finally figured it all out sometime in 2010 or 2011 (remember that he, Crosby, and Staal each missed a large chunk of that season).

With many of the main guns hurt for parts of this year, complementary players like Orpik and Kennedy suffered. I think the key for Pittsburgh, obviously, is the star power being able to carry the team. But the fact that they've let the depth come and goeven on the bluelineI think means that the organization recognizes that star power is harder to come by, and hence more critical to retain, up front. (Think about how defensemen agechances are they're still improving or peaking when they hit UFA, unlike forwards, who are likely declining.) But they've banked on youth maturing to get those green boxes; aside from Bylsma taking over for Therrien, I don't think there's much to glean from this one.

I started this post thinking that personnel changes played a big role in this possession decline. From this short look, here are my takeaways:
  • The Capitals' decline probably started in 2009-2010. I think it was too early for the effects of age to set in, so there must be some other reason. I'd estimate about half of it is related to replacing Fedorov with Brendan Morrison and, later, Tomas Fleischmann, but I'm not sure what that the other half is—it will require a closer look.
  • It continued through 2010-11 until Oates may have reversed it this season. Because basically every player, and some of the team's main duos, saw Corsi declines, I think coaching is the prominent factor here. I don't really see anything to suggest that the system change under Boudreau and the switch from Boudreau to Hunter helped, well, anything, except maybe to stabilize the team in the very short term (while sacrificing the longer term, like the next season). Coaching is also responsible for less-than-ideal deployment strategiesagain, Beagle with Ovechkin—but not injuries, which were likely a factor, but tough to judge as such relative to coaching—Green didn't move the needle as much as he has in the past during his brief "healthy" stints, meaning that he either wasn't himself when in the lineup, or simply wasn't as good as he was in the past. (Or that we got that result through a fluke, which is possible.)
  • With sufficient youth, some utilization changes, or the emergence of players who can dominate minutes (regardless of situation—San Jose needed Logan Couture to dominate easier minutes his rookie season while Thornton took the toughs, for example), the Capitals can reverse some of the decline in their possession numbers. The effects of time, though, are inevitable if the strategy is too much of the latter and not enough of the former.
To end on an optimistic note, the Capitals, in a way, did pretty well in terms of puck possession this year. After basically a) losing Hamrlik, a guy who put up good Corsi ratings under both Boudreau and Hunter but couldn't find a spot in Oates' system and was eventually claimed off waivers by the Rangers, and b) losing Semin, a guy who beat easy minutes pretty handily in the regular season, and c) replacing Semin with a guy who, at best, was going to break even at 5v5 in Ribeiro, and d) losing a good young defenseman in Dmitri Orlov thanks to injury issues and poor play upon his return, the Capitals were poised for a possession disaster. And yeah, that's what it looked like early on.

But they were a ~50% Corsi tied team over the last 30 games. Having learned the system a little better, and with a return to form from Orlov and a better ES player on the second line (Mathieu Perreault should be better than Ribeiro here, although he may not beat up on his opponents as much as he has in the past), plus full seasons from Erat and Laich, along with better chemistry between Green and Alzner and above-average netminding from Braden Holtby...a division title may be too much to ask for, at least as long as Kris Letang is still helping the Penguins' transition game and free agency and compliance buyouts still haven't changed the landscape, but an East mid-playoff team with the pieces to make a deep playoff run? I'll be suckered into believing this could be the year yet again.

No comments:

Post a Comment