• Jake v. Randy II

    by  • May 13, 2014 • Hockey • 4 Comments

    I wrote a post yesterday about why the Maple Leafs did so much better following neutral zone faceoff wins after the Olympics with Jake Gardiner on the ice than they did when he wasn’t on the ice. It was pretty stark. Post-Olympics, Gardiner was on the ice for 73 NZ wins. The Leafs generated 25 shot attempts for and allowed 13 shot attempts against, a Corsi% of 65.8%. When he wasn’t on the ice (128 NZ wins), they generated 25 shot attempts for and allowed 38 shot attempts against, a Corsi% of 39.7%. The league average in these situations was 59.2%.

    In my last post on this, I explored the difference in how the Leafs were entering the offensive zone and how the opposition was entering the Leafs zone with Gardiner on the ice as compared to when he wasn’t on the ice. We found what we’d expect: the Leafs were more likely to carry the puck into the offensive zone when Gardiner was on the ice. Intriguingly, we also found that the Leafs were less likely to allow a carry or dump-in to their end when Gardiner wasn’t on the ice but more likely to have the opposition botch the entry altogether.

    We care about zone entries because the work of Eric Tulsky and others has shown us that teams get and allow more shot attempts when the puck is carried in rather than dumped in. Today I’m going to talk about the shot attempts. As a general rule, Eric’s found that teams average 0.56 Fenwick events (shots on goal and missed shots) per carry and 0.25 Fenwick events per dump-in. I’m using Corsi, so I’m going to make a back of the envelope adjustment to these figures: 0.75 shot attempts for per carry and 0.34 shot attempts per dump-in. What do we see?

    You have to keep in mind here that the samples are very small. That being said, the difference between Gardiner and the Leafs without Gardiner is tiny for everything other than carries with Gardiner on the ice – it’s a couple of shot attempts in each case. I wouldn’t really read anything into it other than chance. In fact, if we look at the earlier part of the season, we see that while the Leafs did better following NZ wins when Gardiner was on the ice, the Corsi% difference was much smaller: 60% to 54.6%. While I’m inclined to think that Gardiner does do better things following NZ wins than the Leafs do without him, I doubt the difference is as large as it was post-Olympics.

    The gap between the shots generated off of carries with Gardiner on the ice versus when he isn’t quite so small. It’s large enough that I’d wonder about it if I was in Leafs management/coaching, particularly if it had persisted throughout the year. In particular, I’d wonder if the carries that were being created when Gardiner was on the ice were somehow different than the carries created when he wasn’t. It’s possible that it’s chance but, as that gap increases, we should wonder more.

    There’s an important general truth here: the results that we observe in hockey, in anything, are a mix of the underlying talent and chance. The smaller the sample, the more that chance plays a role. I suspect that, while there are real differences between how the Leafs play with Gardiner on the ice and without Gardiner on the ice, there was an element of chance making the gap appear larger than it was.

    I put together a little video to kind of illustrate what I see in being a difference in how Gardiner and Cody Franson moved the puck out versus the rest of the Leafs D. I just used some clean faceoff wins outside the defensive zone from the last ten games of the year or so. What I was looking for was how the puck was handled in the defensive zone and how it was moved out.

    In general, the Gardiner/Franson pairing seemed more comfortable making a few passes in order to create some space in which to skate. The other guys were much more “You make one pass then fire it off the guy on the other side of the red line.” This doesn’t always lead to a carry into the offensive zone but it does make it easier for the defenceman to gain the red line. The Leafs iced the puck 40 times more than the average team at 5v5 this year. Undoubtedly, some of that was defencemen missing sticks trying to run a tip play into the offensive zone.

    I thought the last one on there, with Gardiner/Phaneuf on the ice, was telling. Phaneuf had room to skate the puck after Gardiner did his little shimmy but didn’t – just bombed it towards a stick on the other team’s side of centre. There’s another example of that in there, where Phaneuf could have stepped up but didn’t. I strongly suspect that the bloodbath that occurs when Phaneuf is on the ice these days is coaching – if he has the ability to take that step but doesn’t, it seems to me like an awfully crazy thing. To me, I’d want to identify the things that were working and then ask those with the ability to do it to do that.

    This is, of course, just a tiny slice of the game. This is the sort of stuff that the Leafs should be doing though, breaking the game down into tiny pieces and examining why they did so much better with Gardiner on the ice than everyone else. The big reward is if they can figure out the open play question – Gardiner was ten points better than the team as a whole post-Olympic break.

    James Mirtle has a great interview up with former Leafs assistant Scott Gordon.

    “We have a lot of guys that can make plays and there’s a confidence that [they have] when things are going well that you think you can make plays all the time. At the end of the day, there’s a time and place to do it. We actually got better as the season went on as far as turnovers. But it’s not just the turnovers. You’ve got to find a way to maintain speed through the neutral zone to get on pucks in the offensive zone when you are dumping it in.

    Gordon goes on to talk about how this year’s team didn’t compete hard enough, which, honestly, I just kind of shake my head at. Did they compete harder when Gardiner was on the ice? Or did they do less stuff that has bad expected outcomes? Watching these videos (and looking at the data for with Gardiner on-ice as opposed to when he wasn’t), it doesn’t seem to me like they were trying to create plays. It seems to me like they were obsessed with avoiding turnovers and went about attacking them in a way that had all sorts of negative consequences. Forest. Trees.

    Email Tyler Dellow at tyler@mc79hockey.com


    4 Responses to Jake v. Randy II

    1. john
      May 13, 2014 at

      You can imagine this sort of thing spontaneously evolving based on what players have incentives to do — so even before a “get it deep” coaching strategy is widely accepted, it could become folk wisdom at the player level.

      Players don’t try to avoid making mistakes per se, because to even know what a mistake *objectively* is, you need a worked-out idea of what kind of entries are good, which kinds of turnovers are good (if any), what kind of big-picture results you can expect from individual actions.

      So players do what they can do with their subjective knowledge, which is avoid the *appearance* of a mistake, because appearing to make a mistake gets you benched. So even if hammering it down the other end of the ice is objectively a mistake because it’s a turnover, it looks intentional, and isn’t immediately dangerous, and anyway if someone hustles hard enough they might get it back, so your coach can’t come down on you too hard about it.

      Then, years later, when they’re in coaching jobs, the same guys who learned (even if they were never told explicitly, which of course in real life they were) to “just get it deep” can’t differentiate between what kept them out of their coach’s doghouse when they were playing and what actually leads to desirable outcomes on a team-wide tactical level.

    2. Skinny
      May 13, 2014 at

      Seems like the Leafs’ obsession with limiting turnovers, much like the Oilers’ obsession with winning faceoffs, ended up with them missing the forest for the trees.

    3. Great Dane
      May 14, 2014 at

      Tyler, really interesting stuff.

      After reading the interview with Gordon, it seems pretty clear to me that the coaching staff realized that the Leafs had a number of structural issues with the team. However, their analysis of the problem came to the wrong conclusion:

      The huge number of SA is a problem; we turn over the puck too much so instead of turning it over we dump it in.

      I am sure that we have seen a number of situation where a turnover resulted in a goal against. That doesn’t take into account the second and third derived of a dump in.

      Tulsky has shown that a dump in creates shots on the goal, but what are the consequences of a missing shot, i.e. the second and third derived:

      The other team immediately has the puck and can create a play. Leafs are then in forward movement to chase the puck, while the pucks is going the other given a scoring opportunity to other team.

      Leafs have a lot of hit and as Hitchcock put you only have a lot of hits if you don’t have the puck.

      As you mentioned the other team has more difficulties entering our zone with Gardiner on ice. That shouldn’t be a surprise since Leafs are creating more with Gardiner on the ice than without him. When Leafs have the initiative and pressure the other team it is also easier to defend.

      Would it be possible to compare the Leafs D-men making the same analysis as for Gardiner?

    4. jvuc
      May 14, 2014 at

      Interesting insight but it doesn’t capture everything. When you dump the puck in, the goal is to do so on the team’s weak side (ie. away from speedy dman or towards a dman play on their wrong side). The better teams do employ this when they dump the puck in.

      Secondly, the leafs don’t have enough dmen who can carry the puck in. Gardiner, Kadri and Kessel for sure but after that the quality of carry ins decreases.

    Leave a Reply

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