Week 13 Snapshot: The Unsustainable Lifestyle

8 Patrick McDermott

Photo: Patrick McDermott

In 2004, the DC band Beauty Pill released an album called The Unsustainable Lifestyle on Dischord. It’s a great record, but it can tend to bum you out. The band seems to be aware of all this stuff wrong with the world that you might otherwise not be aware of. If you don’t listen to the album, you might think everything is okay.

So feel free to skip this week’s snapshot.

The Capitals played only two games this week thanks to the holiday. They split the pair, but they could have won both had they not given up a two-goal lead to Anaheim on the 23rd. That’s the party line at least. In truth the Capitals got dramatically outplayed in both games they played. That they held a temporary lead over the Ducks and snuck out with a win over the Rangers is thanks to some of the most unreliable stats in our sport: shooting percentage and save percentage. The more stable numbers describe a Capitals team that is about to lose games– a lot of of them, and badly– unless something changes.

These are the numbers as of noon on Sunday, December 29th. Our sample is 5-on-5 play while the score is close. That means within one goal in the first two periods and tied in the third. That way special teams, blowouts, and comebacks don’t color the data. Stats of note are highlighted in powderpuff pink and discussed below.

See previous snapshots: week 1, week 2, week 3, week 4, week 5, week 6, week 7, week 8, week 9, week 10, week 11, week 12

But first:

Forwards

Player Pos GP TOI GF GA SA% Sh% Sv% PDO ZS%
Brooks Laich C 26 201.4 4 8 40.4% 5.5% 93.2% 98.6 48.4%
Marcus Johansson C 38 322.2 9 16 48.4% 6.1% 90.5% 96.6 54.5%
Mikhail Grabovski C 36 294.7 15 11 47.5% 10.3% 93.5% 103.7 49.8%
Nicklas Backstrom C 38 337.7 11 15 48.2% 6.2% 92.4% 98.6 57.2%
Alex Ovechkin L 36 331.7 7 15 48.7% 4% 92.3% 96.3 58.7%
Jason Chimera L 38 304.9 15 17 43.8% 10.2% 90.8% 101 43.2%
Eric Fehr R 29 221.1 8 11 48.8% 7.9% 90.4% 98.4 48.2%
Joel Ward R 38 291 12 16 46.7% 8.5% 90.3% 98.8 44.8%
Martin Erat R 35 252.2 10 10 47.6% 8.7% 93% 101.7 48.2%
Tom Wilson R 38 168.4 6 5 39.1% 10.7% 94.3% 105 54.6%
Troy Brouwer R 38 289.8 10 9 42.3% 9.3% 94.4% 103.8 50%
Michael Latta C 17 67.6 3 4 48.1% 10.3% 88.2% 98.6 48.5%
Aaron Volpatti L 30 137.6 3 4 34.6% 6.7% 94.7% 101.3 48.6%
Jay Beagle C 18 104.1 1 2 36.8% 3.2% 96.2% 99.5 50%

Defense

Player Pos GP TOI GF GA SA% Sh% Sv% PDO ZS%
Nate Schmidt D 28 270.8 12 10 51.4% 8.6% 93.1% 101.6 55.3%
John Erskine D 13 129.9 6 7 46.7% 10.3% 90% 100.3 44%
John Carlson D 38 383.4 12 22 42.6% 7.1% 90.4% 97.5 48.6%
Karl Alzner D 38 392.8 12 15 43.2% 7.1% 93.1% 100.3 49.2%
Mike Green D 35 395.1 10 13 49.3% 5.2% 93.7% 98.9 55.3%
Steve Oleksy D 29 247.7 16 9 46% 14.5% 93.2% 107.8 49.5%
Dmitry Orlov D 12 136.8 2 6 43.5% 3.5% 92.2% 95.7 43.2%

Observations

  • Despite playing only two games, the Capitals’ puck possession (i.e. even-strength shot-attempt percentage) dropped hard this week. Last week that number was 47.11%. Now it’s 46.30%. There are only five teams with a lower possession score, and four of them are laughing stocks: Oilers, Flames, Leafs, and Sabres. We’ll see that last one in action later today.
  • The only reason the Caps had leads to blow in the Ducks and Devils games was their high on-ice shooting percentage: lots of goals on not a lot of shots. A high percentage like that cannot maintain through the season. The Caps have to maximize the volume of shots, or they will end up this season’s version of the Wild or the Leafs, stubbornly declaiming the strength of their shot quality as their postseason chances evaporate.
  • Same is true at the other end of the ice, where the Caps are facing a ton of shots and stopping almost all of them. The team has gotten stellar goaltending from Philipp Grubauer, who absolutely has the hot hand and absolutely should be congratulated for it and absolutely should be sat in favor of Braden Holtby as soon as possible. I mean, guys, there’s an entire Wikipedia page dedicated to the foolishness of thinking this is going to last.

  • The only players to see their possession numbers improve this week were the top line– Alex Ovechkin, Marcus Johansson, and awesome Nick Backstrom– and Brooks Laich, whose score was so low that ticking up doesn’t mean much. It’s obvious Adam Oates has done much to improve the game of his star pupil, but the rest of the team is really suffering.
  • Here’s the part where I praise George McPhee. Martin Erat and Mikhail Grabovski have, in my opinion, done much to counteract the systemic damage to the team’s puck possession. Those players were GMGM’s marquee moves in the last year, and they’ve both worked despite problems with how they’ve been directed and deployed. Without them, I’d bet the gap between the top line and everyone else would be even bigger. But this is a stat snapshot, so I’ll save further opining for comments and other stories.
  • Aaron Volpatti is not an NHL player. A 40% possession score is usually the lower limit for acceptability; Volpattti is now at 34.6%. That is dead last among all players in the league with at least 21 games played. Thank goodness Caps goalies are stopping 19 of 20 shots he’s allowing while on the ice, but that’s not going to last forever. He plays 0.4% of the team’s power plays and 0.2% of the team’s penalty kills. The puck is in his end more than 65% of the time he’s on the ice in close games. His salary is $575k. What are they waiting for?
  • Volpatti’s most common linemate is Tom Wilson with 165 minutes shared. When they play together, 40.7% of shots (including blocked shots, and not just in close game) belong to the Caps. When they play apart, that number is 29.3% for Volpatti and 53.1% (!!!) for Wilson. Wilson has had a really bad December. He needs a change; here’s the impetus to make one.
  • Eric Fehr: still awesome. He had a grape and an apple this week, and he personally fired six of the team’s 34 shots against Cam Talbot the other night. That was his highest shot total in three years.
  • The other day I was looking at which players had played best with Alex Ovechkin over the last five years. The usual names appeared: Semin and Backstrom, obviously. But down below, among players he shared less ice time with were two names with really good numbers: Eric Fehr and Sergei Fedorov. They only shared 5v5 ice with Ovi for between 180 and 200 minutes, but they owned the puck– 62.2% of shot attempts (again: including blocked shots and regardless of the score) belonged to the Caps. My point: I wouldn’t mind seeing Fehr on the top line.
  • John Erskine is back on the snapshot as he seems to be here for awhile. Gone from the snapshot but not from the payroll is Alexander Urbom, whom I though was waiver-bound a while ago, but oh well. Same old song and dance about the defensive depth here.
  • Still no goals for Martin Erat, though he had an amazing assist the other night. Erat is in the top-20 list of players with a ton of shots but no goals. And despite that, his on-ice shooting percentage in close games is a solid 8.7%, which is surprising.

Glossary

  • GP: Games played
  • TOI: Time on ice
  • GF: Capitals goals for which the player was on the ice
  • GA: Opponent goals for which the player was on the ice
  • SA%: Percentage of shot attempts (from both teams) that went towards the opponent’s net, excluding blocked shots
  • Sh%: Capitals’ shooting percentage while the player was on the ice
  • Sv%: Capitals’ goalie save percentage while the player was on the ice
  • PDO: The sum of Sh% and Sv%, a number that regresses closely to 100 in larger samples; a proxy for luck, in a sense– i.e. high ≈ lucky
  • ZS%: The share of shifts the player started in the offensive zone, excluding neutral-zone starts; data not limited to close games.

Thanks to ExtraSkater.com for the stats.

Tagged with:
 
  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    well that was a bummer

  • Matthew Kory

    From reading these each week, I’ve gleaned that the Caps don’t have the defensive players to make a real run this season and even if they did, their forwards aren’t deployed in anything like an optimal manner. Go Caps Hooray!

  • Michael Reschly

    Is there a word for “Bissonette would be a significant upgrade”

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    haha

  • Red

    Ignorance is bliss. I used to enjoy Caps hockey, but then i took an arrowadvanced stats to the knee. Now I celebrate wins by muttering to myself something about shot attempts and puck possession.

    Also, I believe you’ve made your argument against Grubauer. There’s no point in reiterating it over and over. The stats are in favor of Holtby. However, there’s also the notion of “riding the hot goaltender” and every coach in the league is guilty of that. So lets cheer for Grubby and see how this thing plays out.

  • Ben Reed

    “The more stable numbers describe a Capitals team that is about to lose games– a lot of of them, and badly– unless something changes.”

    Well, one could argue the recent downtick in possession will rise again, coinciding with the regression to the mean of save %, and the Caps will continue humming along as they have.

  • Jimmy Pollock

    Blah blah blah… Week after week after week I read on here that the Caps can’t keep it up because the stats say so. Anyone else notice that the season is basically halfway over now and the Caps DO keep winning games and have remained in 2nd place for some time now? Negative Nancy, maybe your stats aren’t so fancy after all? Last time I checked, the only stat that matters is how many points the team has. What’s next? A “Nicky scoared a goal but it wasn’t pretty enough” article?

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    I don’t think you’ve read carefully then. A month ago, the team’s possession was getting stronger http://www.russianmachineneverbreaks.com/2013/12/08/week-10-snapshot-addition-by-subtraction/

    You’re not appreciating how teams win games in different ways– some by special teams, some by 5v5 possession, some by luck. Usually a combination of the three. What I’m doing is identifying which is driving the team’s fortunes, and them discussing what that means for their future.

    “The Caps DO keep winning games”

    Actually their win % has dipped strongly in the last couple weeks.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    Different stats regress differently. Sh% and Sv% regress way harder than possession.

    Further, there’s little evidence you can drive Sh% and Sv%, whereas there’s a lot of evidence you can drive possession stats by changing lineups and tactics.

  • Ben Reed

    True, but the Caps have been SO bad possession-wise recently, it’s safe to say that is not their “true” possession ability. There is a lot of room there.

    I share your skepticism but just there’s still some water in that glass.

  • JH

    My theory on Grubauer: they lack confidence in Neuvirth’s ability to stay healthy, so they want to give Grubauer NHL reps in case he’s the No.2 going into the playoffs. I don’t take this as a slight vs Holtby.

  • Michael Reschly

    Why do you use Fenwick and not Corsi? Not saying you’re using the wrong one, just trying to educate myself.

    Also, I wonder how well an obsessively fancy-stat optimizing team (shoot as soon as you clear the d-zone, etc) would do in terms of win-loss.

  • Pat Magee

    Maybe we can play the make beliefs in the playoffs first round. That would be great! Nobody would ever have the puck!

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    I use Fenwick because when the sample gets huge, it’s got a slightly higher correlation with goal ratio: http://objectivenhl.blogspot.com/2011/02/shots-fenwick-and-corsi.html

    And also a lot of my favorite data uses it as well (the Habs Eyes on the Prize chart, the BTN possession dude I link to above).

    However, in teeny samples, Corsi is probably better because there are more events sooner. This ain’t perfect.

    I love your second point. I think if a team focused only on maximizing their possession stats, they would lose a lot. Since the point of the game isn’t possession (as measured by shot attempts), but to score goals, the team that just takes a ton of 1% or 2% shots from the blue line– without any rebounds or net-crashing– is gonna get creamed.

    The good news is that nobody does that because it would be silly.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    I hope you’re right! I think they can turn it around.

  • Michael Reschly

    Just like it would be silly to trade MP85 for pucks, put Erat on the fourth line, and F16 on nacho duty?

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    haha well yeah you got me there

    http://www.youtube.com/watch?v=1Npo0cmp-VY

  • time

    the hipster trend of hockey fans continues…do you watch games at the black cat?..wait i guess they don’t have tvs…but its been a while since I’ve been there.

  • 888

    After all, Peter, I think you guys wrote the Caps out of the playoffs last year before they went on their run. (Which, based on their performance and stats at the time, was a totally logical prediction. I’m not arguing that.)
    One of the reasons I enjoy hockey so much is that it really does seem capable of some amazing events to me– some that definitely can be predicted when you analyze the stats (Hello, Sv% and Sh%!) and some that just seem to come out of nowhere, whether we end up referring to it as chemistry or finally getting the system, or whatever.
    It’s still kind of magical and a reason I watch. (Doesn’t have to be the reason anyone watches! Just one of mine.) Because despite the highs and lows, every season, I secretly think, “Okay, this is our year,” no matter what. Eternal hope, and all that.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    ?!

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    Here’s my mea culpa from last season and how the Caps made the playoffs last year: http://www.russianmachineneverbreaks.com/2013/04/24/my-apology-and-how-we-got-here/

    And btw, I completely agree with you on the unpredictability. That’s what makes it fun!

  • tim

    lol…just saying the puck daddy bros get accused on other hockey boards for being “hipsters”…and the beauty pill reference made me think of that.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    The problem with being accused of being a hipster is that you can’t deny it because denying being a hipster is the hallmark of a hipster. Anyways, I used to be a hipster but then it got too mainstream.

  • Marky Narc

    something something old 9:30 club something.

  • riggorules

    I’ve been wondering how taking a lot of low-percentage shots skew possession stats. From my biased eyes (as a Caps fan) I feel there are games where the amount of time the Caps control the puck is not reflected in the shot count. Perhaps they’re being too selective, or not converting good cycle play into attempts. Or maybe they’re just as bad as the #fancystats say and my eyes be lyin’.

  • riggorules

    So is the solution the players or the system? Would be interesting to see a comparison to the Devils, who I believe play the same system. They’re a strong possession team (52.5% FF) with maybe less scoring talent and a worse record. But is there defense better at getting the puck out of the zone cleaner, an apparent weakness of the Caps? If Oates optimizes the lines for possession, is that the difference-maker? Looking for a way out of this trend, or are they just doomed to hope for puck luck.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    I don’t have a great answer for ya! I don’t think the Devils play the same system, but only because the Capitals system isn’t really easy to classify. Caps allow many many more events in both directions.

    There are personnel problems (Volpatti) and there are deployment problems (Wilson) and there are tactical problems (5-foot rule for D) and there are systems problems (dump and chase). It’s tough!

  • Matt Lauer

    I’m going out on a limb and arguing that the shots-against stats are deceptive. I don’t think that nearly as many of the shots the goalies see are the glorious scoring chances they’re made out to be (some, if not a few more than some, are glorious of course — but do you expect professional athletes not to score goals against their opponents?). The Rangers came was a case-in-point. Many, many, many of the Rangers’ shots were from the perimeter and were cleared from the crease area after the initial rebound. It seems to me that the Caps are trying to play that sort of game: give ‘em the perimeter shots, clog the middle, carte blanche in the offensive zone. It’s not entirely unlike the 1998 Finals team.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    Shot quality doesn’t really skew possession stats– particularly close-score ones.

  • JenniferH

    I hope that’s what is going on, I really do, because, damn, I REALLY miss Holts.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    This has been looked into. In bigger samples, scoring chances correlate with possession. That’s why no one really tracks scoring chances anymore.

  • JessHughes

    I’ve seen RMNB reference the “5 FT rule” several times now. To what are you referring and what is your source? Great work on the Snapshot, as always. Go Caps!

  • http://www.russianmachineneverbreaks.com/ Peter Hassett
  • Joe LaCour

    Trade Holtby, Erat, and #1 to Buffalo for Miller. Then trade Neuvirth to somebody for a high #2.

  • Matt Lauer

    Fair enough, though it seems the most useful comparison, in that case, would be a team-by-team analysis of possession vs. shooting percentage (median, mean, or whatever on the latter part). Is it a greater indicator of success to have a set of lines with killer possession but weak shooting, or average possession and strong shooting? Think of the Konowalchuk-Dahlen-Halpern line from all those years back — insane, ridiculous, crazy possession time, but generally poor shooting.

  • http://www.russianmachineneverbreaks.com/ Peter Hassett

    I’d say nearly all analyses have concluded that most fluctuations in team-level shooting % are due to random variance, and can pretty much be ignored when doing predictions.

  • Matt Lauer

    Touché, Hassett.

  • JessHughes

    Peter — thanks! I’m out of my depth here, obviously, but it seems that this explains why our D-men always look like they are playing “hot potato” with the puck. I’d prefer giving more autonomy to our good skating D-men (Green, Dima) to let them skate into the neutral zone with possession.