• Bruce Boudreau and Offensive Zone Faceoff Losses

    by  • April 13, 2014 • Hockey • 8 Comments

    One of the best pieces of data driven hockey journalism that I’ve ever read is J.P’s “When Everything Changed…And If It Needed To.” He discusses, with reference to data, how the Bruce Boudreau’s decision to implement the trap was a disaster for the Capitals.

    As it so happens, I don’t think Boudreau gets enough credit for being an excellent coach. I’m working on something showing how things changed when Randy Carlyle was fired in Anaheim and then how they changed in Toronto when he was hired there. In the course of doing this, I tweeted out something showing how Boudreau’s Ducks and Carlyle’s Ducks had done following faceoffs in certain areas of the ice (offensive zone wins/losses, neutral zone wins/losses and defensive zone wins/losses.)

    Darryl, from Extra Skater pointed out something interesting to me: Boudreau’s teams in Anaheim have had a pretty small difference between their Corsi% following an offensive zone win and following an offensive zone loss.

    This got me thinking. The normal difference is about 20 points. If you win an offensive zone faceoff, your Corsi% over the next 37 seconds will be about 74.5%. If you lose that faceoff, you’ll do the next 21 seconds at about 54.3%. So there’s a sizeable gap in there. That’s the value that winning the faceoff gives you.

    I haven’t run this for this year but Anaheim’s numbers from the time that Boudreau got hired through the end of the 2012-13 season were pretty wild. I’ll include the Ducks’ numbers under Carlyle as a contrast.

    So you can see that the Ducks have done slightly better following offensive zone wins but that the numbers following offensive zone losses are just out of this world, at least through the end of last season. 67% is a crazy good number to post following an offensive zone loss though – so good that I actually went and re-did my math because I didn’t believe it.

    This got me thinking about something. In that Japers Rink piece I mentioned above, the working theory is that Boudreau’s decision to try and turn the Washington Capitals into a trap team doomed him. You really should go and read it but basically J.P. lays out a compelling case that the change in tactics hurt the Capitals and that they then went downhill, Boudreau got fired and, outside of that cool PP trick that Ovechkin does, the team’s pretty much a disaster.

    “Self,” I wondered, “maybe if you back and look at the numbers that Boudreau’s Caps put up after offensive zone faceoff losses, you’ll find something cool. Separate the numbers from Boudreau’s Rock and Roll Period in 2010-11 from those that the Caps posted during his Trench War Phase in 2010-11. Is it possible that Boudreau’s got a track record of teams that excel at this?”

    Hmm. Isn’t this awesome? Basically, every Boudreau team is above average after losing offensive zone faceoffs – most are well above average – until Defence Wins Boudreau comes to town. If you think about what it means installing the trap, it was probably tied to other defensive changes, like retreating towards the neutral zone once possession was lost in the offensive zone – like, say, after an offensive zone faceoff loss. Sure enough, there it is.

    The Caps are right back up the following season. I don’t know if he changed what he was doing or if it’s a small sample size issue but then he gets fired, goes to Anaheim and Anaheim perks right up after they lose an offensive zone faceoff.

    I don’t believe too much in motivational magic, so I would guess that Boudreau’s got a better set of tactics for what to do when you lose an offensive zone faceoff than most do. Whatever he has his teams do (save for a brief period when his faith was tested in 2011-12), it works – his teams are routinely excellent at it.

    The cool thing about this is that it ties into what J.P. observed in his piece. If we had a better data base of tactics and what teams are employing what at any given time, we’d probably be able to make a much cleaner link. As it stands though, I’m pretty comfortable with the belief that what we’re seeing here is the impact of tactics and that Boudreau’s much closer to whatever the analytically correct answer might be than Carlyle is (or his successors in Washington.)

    Email Tyler Dellow at tyler@mc79hockey.com

    About

    8 Responses to Bruce Boudreau and Offensive Zone Faceoff Losses

    1. Woodguy
      April 13, 2014 at

      Wow.

      Those are pretty unreal number after a face-off loss.

      This just adds to the growing evidence that “safe hockey” isn’t very safe at all.

      It seems to reduce your corsi events for as much, if not more than you save in corsi events against.

      I would even go as far to say that “safe hockey” tends to reduce the amount of corsi events and leaves you exposed to whims of variance in SH% and SV% as well. NJD is an example of this.

      Paints a grim future for the “safe hockey” that MacT and Eakins subscribe to in terms of what we can look forward to in terms of the Oilers.

      Sacrificing corsi events for on the alter of avoiding the boogeyman of blue line turn overs helped drive Hall’s into the ditch.

      Sacrificing corsi events for on the alter of a losing streak due mostly to poor SV% (As JR mentions in his excellent piece) drove WAS into the ditch.

      I can see why Eakins may want to preach and get the team to buy into safe hockey, they don’t have the roster trade chances (when Nick Shultz lines up for >50% of the nights as 2LD, you don’t have a roster for doing much), but what is lost in the process?

      Have you watched a bunch of ANA games to see what Boudreau is doing?

    2. Bling
      April 14, 2014 at

      This reminds me a bit of piece I read on the effect that analytics was having on basketball.

      One of the conclusions was that teams ought to be much, much more aggressive on help defence than they currently are. Interestingly, very few coaches were interested, the reason being that more help defence = more dunks for your opponent = more of you “looking bad”, even though over the long term your team would be more successful.

      Maybe that’s what happened with Boudreau in Washington. He was having great results, but there was an air of defensive indifference — a few odd-man rushes the other way, a few GA that “looked bad”. So they flip-flopped.

    3. Bling
      April 14, 2014 at

      Also:

      All of Mike Green’s monster offensive seasons coincided with Boudreau’s “Rock N Roll” era.

      Makes me wonder if the Caps were more aggressive than other NHL teams in activating the defence to maintain possession following an o-zone draw, regardless of outcome.

    4. Jeremy Wright
      April 14, 2014 at

      Could this partly be an effect of playing in the SE division? You should break it out into non SE opponents and see what happens.

      Great stuff as always:)

      • Jeremy Wright
        April 14, 2014 at

        And I should clarify… obviously BB is doing something unique, but how much of the Washington inflation was due to playing in the SE?

    5. April 14, 2014 at

      Jeremy it is a valid argument that the Caps benefited from weaker teams in the Southeast. However, BB had his teams marching over everyone in the East in 09-10. Also, he is no longer getting the benefit of those weaker teams out West. The sample size is now large enough to show that it doesn’t matter who the competion is, BB has an effective system in place to generate offense even after losing a face off.

    6. FastOil
      April 14, 2014 at

      Interesting piece as usual, thanks.

      I wonder how could his tactics be a secret to those with deep hockey knowledge? Is it coaching players with all world talent that play a ton and dominate possession? Why wouldn’t such success be copied?

    7. Rudie
      April 15, 2014 at

      You mention using a window of 37s after OZ+ yet 21s after OZ-. Why the diffence? Do the tables also use the same times?

    Leave a Reply

    Your email address will not be published. Required fields are marked *