• The Details of the Devils’ Difference

    by  • May 6, 2014 • Hockey • 7 Comments

    Nothing in this series of posts identifying differences between the Devils and Flyers after defensive zone faceoff wins would be possible without the fantastic work that Corey Sznajder has done counting zone entries. He aims to finish the entire NHL by the summer. Please consider a donation to support him.

    After I wrote my post yesterday identifying areas in which the Devils produced markedly fewer shot attempts for and allowed markedly fewer shot atempts against following a defensive zone faceoff win, I got to thinking about it a bit. One of the big differences that we saw was in terms of shot attempts generated from pucks carried into the offensive zone. Philly generates lots of shot attempts off those carries; the Devils are generating nothing.

    I did what any sensible person would do and grabbed all the video of Devils and Flyers controlled entries within the 21 seconds following a defensive zone win that was identifiable from Corey’s data. I’m still figuring how best to present video – it’s an internal battle for me between “Show your work because you can ‘prove’ anything with video” and “Don’t dump 20 minutes of raw video of carries on people.”

    This is a 15 second example of what I think I see more frequently with the Devils than with Philadelphia. Basically, it looks to me like New Jersey are far less likely to send a third man up the ice to support the carries.

    Here’s another example with New Jersey, this time illustrated with a series of photos.

    Note Peter Harrold (10), in front of the Devils’ net.

    The Devils win the puck up the boards and force it to Travis Zajac. Zajac passes it to Bernier. You can see Peter Harrold coming hard. The Devils have the makings of a 3 on 2 or a 4 on 3 here.

    Except that Zajac doesn’t even skate. He just kind of rolls up the boards and cruises through the neutral zone. I see this enough in the videos that I’m suspicious that it’s coached. Oh well, at least Peter Harrold is there, right?

    Except Harrold just peels off. So New Jersey is only sending two guys in on what could have been on a three on two. It’s not late in the game or anything.

    Having watched the Flyers and Devils, it feels like I see this more in the Devils games. Essentially, two guys are pushing the puck up the ice. A third might come as far as the blue line but he really requires some obvious chance before committing.

    I don’t like working just off video. My brain’s as susceptible to getting fooled by a narrative that ties nicely into what I already suspect as anyone else’s. So I try and ask myself: “If what I think was true, what else would be true?” I’ve got a suspicion that New Jersey stink at generating shot attempts off carries in the sample that I’m looking at because they don’t throw a lot of bodies at it, which makes it easier to defend.

    If New Jersey is holding someone back to prepare for when the puck comes back into the neutral zone while the Flyers are more inclined to attack with three guys, wouldn’t we expect New Jersey to do much better than the Flyers when the puck came back into the neutral zone?

    As it so happens, with Corey’s data, I can address this question. This stuff is all sequential. There’s a DZF win. The puck goes the neutral zone. One of the teams moves it into the offensive zone. Neutral zone.

    So, for example, the Devils played the Oilers in a game in October. With 1:03 left in the first, the Devils won a defensive zone faceoff. Four seconds later, they dumped the puck into the Oilers end. With 48 seconds left to go, the Oilers carried the puck into the Oilers end. What I can do is number all these entries: the DZF is 1, the next entry 2, etc.

    So I can ask myself who’s more likely to get the next entry in Devils and Philly games. Here’s the table:

    Obviously, the DZ faceoff is 1 each time. New Jersey got the second entry 79 times and allowed the second entry 32 times. This makes sense from a hockey perspective – you move the puck up the ice and there’s risk that you’ll get into the neutral zone and turn the puck over before entering the opposition’s end of the ice. For Philly, it was 85 for, 46 against. So New Jersey’s got an edge there in terms of winning the next zone entry, 71% to 65%.

    What about the next entry? Well, there’s usually not a next entry in the 21 second window that I’m looking at. There’s a big edge for New Jersey though in terms of winning that next entry: they got 10 and allowed 12. Philly got just 7 and allowed 20. 45% to 26%, although sample size concerns abound. It’s fascinating though because if anything we’d probably expect New Jersey to have been slightly worse, given that the puck was more likely to be coming out of the offensive zone for them than it was for Philly, given the results on the second entry. We knew, from yesterday’s post, that New Jersey had an edge on the entries post-DZ faceoff win; this is just kind of showing how it breaks down.

    Taken as a whole, I think that this provides some support for my theory that the Devils are making a bit of a tradeoff here: they don’t push a third player forward when they emerge from their zone and attack the other end, which reduces the shot volume that they generate but puts them in a better position to win the next entry when the puck comes back into the neutral zone. Essentially, they’re more likely to have three guys behind the puck instead of two.

    Two other interesting points. I broke the opposition entries number 2 and 3 into carries and dump-ins.

    If the entry was the first entry (“2″ in my table) after a defensive zone faceoff win (excluding the DZ faceoff), New Jersey were slightly more likely to force the puck to be dumped back in. Both teams were more likely to force the puck to be dumped in on the first entry after a defensive zone faceoff win than they were if it was the second entry (“3″) after a defensive zone faceoff win. New Jersey did better at forcing dump-ins on the second entry than the Flyers did, which I think is consistent with my theory about the Devils keeping more men behind the puck.

    The difference between the first and second entry, in terms of more dump-ins being forced the first time, makes a lot of sense if you translate it to hockey in your head. If you win a draw, get to the neutral zone and lose the puck, you’re probably going to have four or five men behind the puck. If you’ve entered the opposition’s zone and then lost the puck, there’s probably going to be more room in the neutral zone for the other team to make a play, making it harder to force a dump-in.

    I’m still going to go through the video on this but I thought it would be useful to illustrate how the data can be used to check whether what I think I’m seeing in the video makes sense. I can, and probably will, expand on this by going beyond my 21 second window at some point. The possibilities are pretty much endless, once the data exists.

    Email Tyler Dellow at tyler@mc79hockey.com


    7 Responses to The Details of the Devils’ Difference

    1. May 6, 2014 at

      This is rather intriguing and it would help explain why the Devils aren’t shooting so much. Their conservative approach would lend itself to that. Were they mostly a dump-and-change or did the same five stay out there? Given how often you saw it, I’d figure not but that may be something worth noting too?

      I also wonder if they do it as a way to “reset” themselves; intentionally trying to force a dump-in and then trying to move it forward as a response. They limit themselves but the intent may be to limit the opposition more.

    2. Trentent
      May 6, 2014 at

      I would hypothesize that 3 on 2 off a DZ face-off win has a low scoring percentage (rested players, players able to come back hard). As time passes on a shift you are probably less likely to recover from mistakes as fatigue will take its toll (your video took about 8 seconds for NJ to make the OZ, about 25% of an average shift length [not validated]).

      Probably not explained well.

    3. chelch
      May 7, 2014 at

      This is really interesting.

      My question is: If the Devils are the only team playing like this, and it’s not working, shouldn’t they stop playing like this?

      Or maybe a more important question no one is asking – maybe because they’re afraid to ask it – but should Lou Lamoriello retire?

      Can your data support a case for this? or against it?

    4. Triumph
      May 7, 2014 at

      This is really good work. If it’s borne out that they’ve been doing this since DeBoer got here it’s no wonder why Ilya Kovalchuk’s numbers took a bit of a tumble – this kind of system really doesn’t make much sense for his game, which is a counter-attack, rush heavy strategy. Get the puck to the left boards, pull up if the opposition has the middle covered, and look for the late man. Obviously this strategy has all sorts of issues, the first of which is that counter-attack upon failure of this plan is really easy.

      Thing is, now that the Devils have an elite goaltender, I’m not entirely sure that playing low-event hockey like this makes sense.

    5. Jerod
      May 10, 2014 at
      • Tyler Dellow
        May 10, 2014 at

        Do you have a point?

    6. May 19, 2014 at

      This is great work, and corroborates my impressions from watching NJ’s games these past couple seasons: when they dump in, it seems to be just a single player chasing; and when they carry in, the player on the puck gets ridden out to the boards, and either loses it or passes to a defenseman for a long-range shot.

      One question: do you have enough data to do home/road splits for these numbers? NJ’s narcoleptic home scorer makes shot-volume numbers in that building tougher to interpret, but if you’ve got enough games it might be worth seeing how their road numbers compare with Philly’s road entries/shots.

    Leave a Reply

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