• Jake v. Randy I

    by  • May 13, 2014 • Hockey • 6 Comments

    If you follow the Toronto Maple Leafs – and if you live in Toronto, you can’t really help it – you’ll know that one of the intriguing stories about the team over the past two seasons is an apparent disconnect between Jake Gardiner and Randy Carlyle. Carlyle seems to feel that Gardiner makes too many mistakes and that he needs to simplify his game, which seems to be code for more off the glass and out, more tip play dump-ins. More safe play, less risk.

    The curious thing about this, from an analytics perspective, is that the Leafs have had a notably better Corsi% over the past two season with Gardiner on the ice than they’ve had with their other defenceman (min. 1000 minutes at 5v5) on the ice. I’ve summarized those results, along with the goals for/goals against.

    The gap between Gardiner and Gunnarson/Phaneuf is pretty big: five points of Corsi% is huge. Franson’s pretty close to Gardiner, although both Franson and Rielly do better with Gardiner than they do without him. The only reason that Gardiner’s not the Leafs runaway GF% leader at 5v5 amongst defencemen over the past two years is that the Leafs have a lousy S% with him on the ice.

    Since we know that defencemen have a negligible impact on on-ice S%, there’s no reason to expect Gardiner to be so far back of the other four guys going forward. If you were doing an expected PDO for these guys going forward on the Leafs, they’d probably all have the exact same number, whatever that might be. This is why his impact on Corsi% is so important for a defenceman.

    I was looking at some numbers for Gardiner and the Leafs post-Olympic break this season and I noticed something interesting. I’ve put it all into a table, along with the league averages, to give it some context.

    Gardiner did a little worse than the team as a whole following defensive zone faceoffs, somewhat better following offensive zone faceoffs and then miles and miles better following neutral zone faceoffs and in open play (when there hasn’t been faceoff for a period of time that varies based on the location and outcome of the faceoff).

    I decided to focus on the neutral zone faceoffs. What is it that happens when Gardiner is on the ice that produced results that were so dramatically different than those achieved by the rest of the team? In order to figure this out, I watched the 201 Maple Leaf neutral zone faceoff wins post-Olympic break. I observed what happened, in terms of zone entries and zone time, after each of these neutral zone faceoff wins, until there was a whistle or 29 seconds had elapsed, whichever came first. (People always ask: the justification for the 29 seconds is found here.)

    Let’s start with something straightforward. There are basically three different types of neutral zone faceoff: outside the offensive zone, at centre ice or outside the defensive zone. As you can see, the splits when Gardiner was on the ice or not on the ice were pretty similar. It seems reasonable to conclude that the Leafs weren’t making their decisions in terms of when to put Gardiner out for a draw based on where it was in the neutral zone and that his numbers are substantially similar to those of the team.

    I broke the entries down into four different types: carries, dump-ins, failed entries (basically offsides) and “X”, which covers things like a team bringing the puck back into its own end or botching a tip play or lobbing it out of the defensive zone and having it roll into the offensive zone. As you might guess, I’m also using a slightly different definition of “entry” – it’s more like I’m recording where the puck is on the ice.

    The first table that I’m going to put up here is the raw counts of each type of entry and then the percentage of the whole that that represents.

    So, for example, the Leafs carried the puck in 16 times after a neutral zone with with Gardiner on the ice. That represents 20% of Toronto’s total zone entries with Gardiner on the ice. They carried the puck in 20 times when he wasn’t on the ice, which represents 15.7% of their entries.

    I’ve highlighted the things that leap out at me from this table. (I wrote this sentence before I started highlighting, at which point I realized that I was highlighting more things than I wasn’t highlighting but there’s nothing I can do now). Let’s go through these things.

    First, with Gardiner on the ice, the Leafs were more likely to carry the puck into the offensive zone and less likely to dump it in. We know, from work that Eric Tulsky’s done, that that tends to result in more shot attempts. As we’ll see in my next post on this, that rule holds true in this case.

    Second, Toronto’s number of “X” zone entries is higher with Gardiner on the ice than it is when he isn’t. I suspect that that’s the result of blown passes in an attempt to generate a carry but I’d have to go back and review the video to be sure. Even if it is, it seems like a price worth paying to me, subject to the possibility that Gardiner’s generating an extraordinary amount of icings. Generating successful carries is important and the Leafs generate a lot more of them when Gardiner’s on the ice.

    Turning to the defensive side of things, it’s intriguing to me that the opposition gets fewer carries AND dump-ins when Gardiner’s on the ice. The lower numbers there explain the higher number of “X” entries when Gardiner’s on the ice than when he isn’t. As you’ll recall, I include a bunch of different things in “X”, including plays where the puck is drawn into the zone off of the faceoff. There’s something intriguing going on here though: only 23% of the “X” entries when Gardiner’s not on the ice are something other than pucks being drawn into the Leafs zone. When he’s on the ice, 40% of the “X” entries are something other than pucks being drawn into the zone.

    This suggests that the Leafs might be better at forcing the opposition into bad entries when Gardiner’s on the ice (or that Gardiner enjoyed some good fortune). Obviously, that’s going to pay some dividends in terms of driving down the rate at which you allow shot attempts – forcing the opposition to botch an entry beats letting them skate into your end with the puck. It’s probably worth taking a closer look at.

    Taken as a whole, this data enables us to comment a little more intelligently about why things were different after NZ wins when Gardiner was on the ice post-Olympic break. Quite simply, the Leafs carried the puck in more and allowed fewer carries into their end of the ice. These are, obviously, good things. I’m going to write about a couple of other aspects of this – the shot attempts in certain circumstances and zone time – but, on a preliminary basis, the underlying data suggests that there are easily understandable reasons that the Leafs succeeded with Gardiner on the ice and that he might have a case that the coaches ought to give him a longer leash. Moreover, it might be worth asking whether the non-Gardiner defencemen are missing opportunities to generate carries as opposed to dump-ins.

    I’ll leave this here. In the next post, we’ll look at the shots generated off entries and zone time.

    Email Tyler Dellow at tyler@mc79hockey.com


    6 Responses to Jake v. Randy I

    1. Saj
      May 13, 2014 at

      Good stuff as usual.

      One thing I don’t understand. If this is about draws the Leafs win in the NZ, why do the opponents seem to be getting more zone time in the next 29 seconds? Specifically I’m referring to the 57% to 43% ratio which suggests the opponents are dumping/carrying/X’ing into the Leafs zone by a significant margin even when the Leafs win the draw. Are those reversed by any chance?

      • Tyler Dellow
        May 13, 2014 at

        It’s because of how many times the opposition gets an “entry” off a DZ win. So, for example, Bozak wins the puck back to Gardiner and the other team pressures. I count that as an entry, even though it isn’t really.

    2. Actual Hockey fan
      May 14, 2014 at

      It must be depressing to take such a wonderfully entertaining sport and turn it into an exercise in examining ASCII characters.

      • Asst2RegionalMgr
        May 14, 2014 at

        Not everyone looks at a given thing the same way. I’m an analyzer and this is how someone who thinks like me gets further enjoyment out of the sport. Some want to know the ‘why’ of the event.

      • May 21, 2014 at

        The idea Tyler’s not an “actual hockey fan” is insane.

    3. Jaker
      May 14, 2014 at

      This is good stuff but you’re dealing with very small sample sizes. Why not expand at least Table 2 to include this season and last to see if those trends hold up?

    Leave a Reply

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