• Off The Glass And Out And I’m Sure Everything Else Will Be Fine

    by  • June 4, 2014 • Hockey • 15 Comments

    Well, you haven’t really done anything as a hockey analytics guy until you’ve had a spreadsheet in which you recorded the zone entries and exits in the 21 seconds following 135 Oilers defensive zone faceoffs with Taylor Hall on the ice in 2013-14 blow up and leave you with nothing but your own recollection of what it said. This happened to me a few hours ago. There’s no way I’m going to do the work again because it’s tedious enough doing this sort of entry the first time, which sucks because it was really fascinating stuff, but I can recall a few critical things that will help me get at least one post out of it. C’est la vie.

    As certain people (like me) have beaten in the ground, Taylor Hall had a very good Corsi% from 2011-13 (51.4%) and then his Corsi% stunk this year (44.4%). Some don’t really think this matters because POINTZ! but you can’t show me a contending team with a first line running up a Corsi% like Hall. When the Oilers become a contender, Hall’s going to be putting up a Corsi% north of 53%.

    The question of why he ended this year further from that mark than he did his second and third year in the NHL is a fascinating one to me. It seems to me that there are three options. First, this Oilers team might have been markedly worse than the 2011-13 Oilers. I don’t really believe that to be the case. Look at the roster. A lot of the key guys on the 2013-14 Oilers were the same guys on the 2011-13 Oilers, except closer to the prime of their careers. Second, there could be something physically wrong with Hall. As a proud watcher of the games, I don’t really believe this to be the case. Third, the Oilers made some sort of a change that hurt them when Hall’s out there. I think that this is where the answer is.

    As I’ve mentioned a few times, I like to break things down at 5v5 into different game states based on whether there’s been a faceoff in the preceding X seconds, where it was and who won it. X changes depending on the location and outcome of the last draw. I was slicing and dicing the numbers with the Oilers and Hall and noticed something interesting in the 21 seconds after defensive zone wins. Let’s look at the data for the Oilers in Hall’s career, both with and without him on the ice.

    Screen Shot 2014-06-04 at 1.45.01 AM

    Just for a bit of context, the league average in these situations was about 43% last year, a point lower than the average over the past few years. So the Oilers have stunk at this, except for Renney’s last year (we knew not what we had). Hall’s had an average year, a very good one and then two awful years: his rookie year and last year.

    When I started fiddling with the numbers, I noticed something interesting. Up until Christmas, Hall was doing pretty well following defensive zone faceoff wins. 15 SAF 12 SAA. After Christmas, he got slaughtered: 3 SAF 24 SAA. This got me thinking: if I went through the defensive zone faceoffs in a methodical manner, I could determine how things were different, in terms of zone entries and exits, before and after Christmas.

    So I did it. And then my spreadsheet cratered. So I’m left with the one piece of data that I recall. The one, tantalizing piece of data that suggests that there’s something going on here. In recording the faceoffs, there were two out of 135 that I thought ought to have been recorded as losses so I ignored them. As it so happened, this gave me 67 defensive zone faceoff wins before Christmas and 66 defensive zone faceoff wins afterwards.

    The only bit of data that survived my spreadsheet was a breakdown of the first zone exits following the defensive zone win. They’re highly suggestive. I broke them into three categories: carries, dumpouts/other and Didn’t Clear Defensive Zone in 21 Seconds/Before Whistle.

    Screen Shot 2014-06-04 at 2.03.46 AM

    Basically, what we see here is that after Christmas, the Oilers turned a bunch of zone exits that involved the puck being carried out into instances in which they weren’t able to clear the defensive zone within 21 seconds or before a subsequent whistle. Zone exits are important because exiting the zone with the puck under control increases the chance that you’ll enter the offensive zone with the puck under control, which ups your chance of getting shots. It also reduces your chances of seeing the puck immediately shot back into your end. This seems like a poor trade to me and, as we’ve seen, Hall’s results went in the toilet.

    So why did that happen? Well, I’d love to be able to go back and look at the video and see if there are any common themes but, regrettably, I can’t. I can say that it seemed to me like, after Christmas, the Oilers seemed to run a certain play a lot when they won a defensive zone faceoff and it didn’t seem like it was paying off. Essentially, it boiled down to this: they win the faceoff, Hall takes off for the neutral zone and they try to go off the glass. I noticed it so often that I was taking screen captures and tweeting them with sarcastic commentary.

    Screen Shot 2014-06-03 at 4.24.24 PM

    Screen Shot 2014-06-03 at 3.16.39 PM

    Screen Shot 2014-06-03 at 2.10.38 AM

    Screen Shot 2014-06-03 at 2.01.42 AM

    Screen Shot 2014-06-03 at 1.56.53 AM

    A brief video of three of them:

    It seemed to me like this kind of breakout became more common after Christmas, although I acknowledge that this is anecdotal. Further, it seemed like it wasn’t working a lot, which is consistent with the Oilers getting murdered with Hall on the ice following DZ wins. A guy with the data I had could easily confirm this. I’m skeptical that this is a good way to break the puck out of your end after you win a defensive zone faceoff.

    Just more broadly on that point, I put together a list of how each team in the NHL did in the 21 seconds after winning a 5v5 defensive zone faceoff this year. It’s a mixed bag.

    Screen Shot 2014-06-04 at 2.32.41 AM

    Some of those teams, fine, they’re flat out better than the Oilers. There are lots of poor to average teams who manage to execute in this situation a heck of a lot better than the Oilers though: Florida, Washington, Nashville, Phoenix, Buffalo, Calgary and the Isles all missed the playoffs; all were much, much better than the Oilers in this situation. I have a hard time believing that it’s just an issue of the Oilers players being worse.

    Now, I have no idea what any of these teams are doing so much better than the Oilers. It seems, at the very least, possible that they’re approaching things differently from a tactical perspective. That seems like the most likely case to me, to be honest, because I’m not sure that they have players who are dramatically better than what the Oilers have. It’s an easy enough thing to check: cue up the videos and look at how they’re doing it.

    I’m not certain that my theory’s right as to what, particularly, went wrong. I am reasonably confident that there are fairly straightforward gains to be had with the players that the Oilers have. They aren’t going to turn over 80% of the team this year and they aren’t going to win anything with Hall running at a 44% Corsi%, so it’s something that they’ll need to figure out internally.

    Email Tyler Dellow at tyler@mc79hockey.com


    15 Responses to Off The Glass And Out And I’m Sure Everything Else Will Be Fine

    1. chelch
      June 4, 2014 at

      Didn’t the Oilers acquire Fraser (one of my favourite “puck-off-the-glass-and-out” guys) shortly after Christmas?

      Also, there’s a trend in your posts where you point something out and then allege it may be due to coaching tactics, but you leave it open-ended. I feel like you’re doing so much work on the setup and then nothing on the punch-line.

      • Tyler Dellow
        June 4, 2014 at

        There’s only so much I can say without a lot more work. I’m trying to identify a potential issue.

    2. woodguy
      June 4, 2014 at

      Still yet another data point that “safe hockey” is actually anything but “safe”

      I’m sure MacT was thrilled that Hall is “living to fight another day” rather than getting scoring chances.


    3. Kris
      June 4, 2014 at

      When you say you don’t have access to the video, do you not have gamecenter? Guessing it’s probably not worth the investment given the blackout restrictions in Canada but I’m guessing it’d be nice to have for analysis during the offseason (I like having it for fantasy hockey purposes).

      • Shane
        June 4, 2014 at

        This year I believe the NHL posted all of the regular season games on YouTube.

      • Tyler Dellow
        June 4, 2014 at

        I mean my time stamped log of cases where they turned the puck over without clearing it is no more.

    4. G Money
      June 4, 2014 at

      Very interesting… this actually lines up very nicely with a deconstruction that I did of Yakupov’s season w.r.t. Corsi (http://www.coppernblue.com/2014/5/3/5677352/the-two-weeks-that-wrecked-yak).

      Yak had a mostly decent and positively trending Corsi, culminating in a 100% CF% game in mid-Dec. Just saddled with horrible shooting luck and the occasional bad defensive breakdown for the most part.

      The last two weeks of December – coinciding with the possible tactics change you are postulating is showing up in your data – were an absolute disaster. From January on, Yak’s Corsi was poor and trending slightly downward.

      The team followed much the same trend, though on balance Yak was better-than-team early (+1.42%) and worse-than-team (-3.49%) late. A 5% swing in CorsiRel from mid-December is huge.

      Like you’re seeing with Hall, I’d wager that a change in tactics that results in a move to off-the-glass instead of carry-it type of breakouts will affect the high talent players more.

    5. Saj
      June 4, 2014 at

      I find it difficult to understand your defense of Eakins when you’re able to find instances where he institutes changes that make the team worse. You are a fan of his process, but what kind of process leads you to stick to changes that make your team worse? Shouldn’t he have pivoted to a better tactic? Or does his end goal not use Corsi but rather GF/GA/WINZ!, in which case his process would take several years (which he will not get as a coach!) to bear out?

    6. David
      June 4, 2014 at

      I really wonder about how much Eakins tried to compensate for roster/personnel issues with things like this; he overplayed the top line in terms of minutes, but I believe it was a compensation attempt to ensure the horribly outmatched bottom line only hit the ice for a few minutes a game.

      In the second half of the year, the Oilers gave Eakins defenders like NSchultz, Fraser, a couple rookies, and he’d already acknowledged they didn’t have basic defensive zone knowledge.

      With talent limitations and other stuff to teach, quite possible he didn’t trust his defense to make good breakout plays, and coached them to simply get it out.

      • Tyler Dellow
        June 4, 2014 at

        Yeah, I’m sure that this is part of it. At the same time, the defence in 2011-12 was a horror show. Cam Barker (25 GP), Taylor Chorney (3), Gilbert (47), Peckham (54), Petry (73), Plante (3), Potter (62), Bryan Rodney (1), Nick Schultz (20), Smid (78), Sutton (52), Teubert (24) and Whitney (51).

        Compare that with Belov (57), Fedun (4), Ference (71), Fraser (23), Grebeshkov (7), Hunt (3), Klefbom (17), Larsen (30), Marincin (44), Petry (80), Potter (16), Jultz (74), Nultz (60) and Smid (17).

        The second group seems better to me, particularly after Christmas when there’s a lot of Marincin, Petry, Justin and Ference/Klefbom in the top four. Andy Sutton was a top four regular in 2011-12.

        • David
          June 4, 2014 at

          Quite possible the defence now is actually better than a couple years ago (egads); hard to say much in comparison without similar data and analysis re: postDZWin. It’s also possible Eakins was under so much pressure that he just implemented changes that a more secure coach would not have done.

          I’d also be curious to know if the “off the glass” tactic was used uniformly among all defense or selectively (eg, maybe I’m sort of ok with Fraser or NSchultz chipping instead of trying to carry, but I’d hope very much that JSchultz was carrying). Or is this kind of selective tactical deployment not a normal thing?

        • Robert
          June 5, 2014 at

          You actually found a way to make me appreciate our Dmen this year: Just look at who we had two years ago!
          I can’t wrap my head around how our top 6 are going to look next year, but… it scares me to think about it.

    7. June 4, 2014 at

      I wonder how much of the difference between teams is talent and how much is luck in the current sample. It’s not a huge number of shot attempts. It would be interesting to distribute last season’s games randomly into two buckets (“A” games and “B” games) to see how much difference there is between the two. That cuts the overall sample in half, so that might not be the best method, but it seems possible to me that some bad teams are near the top of the list because the sample is too small to separate the wheat from the chaff.

    8. Dwayne
      June 5, 2014 at

      I think the screen grabs are interesting. In all 5 cases, the defenceman going back to collect the puck is facing the near boards on his forehand with a crowd of oilers and opponents on that side of the ice. The only cases where he might have an open oiler as a potential target are the first picture (where he could lead Eberle with a pass to the open side) and the third where Eberle (or whoever) is alone on the other side of the ice (although the 5th King is offscreen and could be right beside him). But to make those passes, the D would have to make a quick pivot to his backhand and then either pass through the crease or skate behind the net (by which time an opposing Dallas player has probably moved to cover the open man). In other cases, if the D tried to make the same pivot, he’d be instantly swarmed. In picture 2 especially, the only available option is ‘off the glass and out’. I think you’ve made a pretty strong case that there is a tactical explanation for your statistical finding – the alignment and movement of players post-faceoff is resulting in the Oilers having no option but a low percentage Hail Mary (or Hall Mary?) after a supposed face-off ‘win’.
      This reminded me of your post from a little while back about different tactics on offensive faceoff losses and the variable success of certain teams and coaches in winning the puck back and generating shots – sort of the inverse of what is happening with the oilers after defensive faceoff wins.

    9. Stocc
      June 6, 2014 at

      How cool would it be if you slowly and painstakingly collected the data a second time and your spreadsheet crashed again? #OilerRebuild

    Leave a Reply

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