• Oilers’ ES Forward Deployment Down 2-1 To Colorado

    by  • February 4, 2013 • Uncategorized • 6 Comments

    My girlfriend and I like to go sailing once a year for a week. We’ll rent a sailboat and sail around some place. Truth be told, I’m a pretty crappy sailor: I can carry things onto the boat and have the presence of mind to vomit over the side in rough seas rather than onto the floor of the boat. Fortunately, she knows what she’s doing, so it works out.

    When we’re sailing along, she’s constantly fiddling with the sails and the heading (and, more importantly, constantly bugging me to fiddle with various lines), with the aim of making small changes that ever so slightly improve what the boat’s doing. Basically, if you make slight changes in how the sails are positioned, how much sail you have out, the precise heading you’re on, you can add little chunks of speed. An extra tenth of a knot of speed here, an extra tenth of a knot of speed there, it adds up. And it prevents me from reading my book.

    ANYWAY, as I looked at the TOI after the Colorado game and noticed how little TOI at ES the Hemsky/Gagner/Yakupov line had, I kind of wondered what was going on. The table at left shows the ES TOI for the various Oiler forwards, both for the game as a whole and then for the post 2-1 portion of the game, the second period after Colorado scored and the third period. A proposition: the Hemsky/Gagner/Yakupov consists of a guy who is 29, a guy who is 23 and a guy who is 19. They are almost certainly as physically capable of playing a pile of minutes in a hockey game as the Hall/RNH/Eberle line is. In a game like this, where the Oilers are trying to get a goal to tie the game, it’s fair to wonder why they played so much less time.

    Prior to Colorado taking the 2-1 lead, the RNH line had played a collective 29.73 at ES and the Gagner line had played a collective 20.30, approximately 68.3% as much ice time. Afterwards, it was 20.33 for the RNH line and 14.30 for the Gagner line. I don’t really have any difficulty with the assertion that the RNH line is better than the Gagner line at ES (although I’m not sure it’s as good as it could be; I prefer Hemsky to Eberle). It does, however, seem to me that it’s sort of odd that the RNH line gets so much more ES TOI with a goal needed to tie – even if you prefer the RNH line to the Gagner line, which is reasonable, surely Ralph Krueger prefers the Gagner line to the other two lines when it comes to getting a goal at ES and, if they’re physically able of playing as much as the RNH line, one would hope to see them having similar ice time numbers when the Oilers are chasing a game.

    There’s nothing to complain about in how the ice time was allocated after the Avs scored in the second period. The Oilers had just come off a power play, so neither the RNH nor the Gagner lines were really an option. The TV timeout, which I’m going to talk about some more below, came at kind of an awkward place, just after the RNH line had taken the ice. This prevented Krueger from really bumping the RNH line there, with full shifts before and after a two minute TV timeout. Really though, the reason for the RNH line having more TOI in the second after Colorado took the lead is the fact of the RNH line getting an extra shift in before the end of the period.

    This is where I kind of think that Krueger made an error. It’s not a major error but then running the bench of a hockey team strikes me as sort of like being sailing a boat – you’re trying to coax extra tenths of a knot of speed out of the boat by changing the sails. This is where the TV timeout rules come into play. There are two minute TV timeouts at the first whistle after the six minute, ten minute and fourteen minute marks of each period, unless there is a power play, a goal has just been scored, or the stoppage was as a result of an icing.

    It strikes me that, if you’re an NHL coach, it isn’t really a game of three periods – it’s a game of 12 periods. A six minute period followed by a two minute intermission, a four minute period followed by a two minute intermission, a four minute period followed by a two minute intermission and then a six minute period followed by an 18 minute intermission. You then repeat this twice and that’s a hockey game.

    If I’m the coach of a team that’s trying to get back into the game and I have two lines that can be reliably counted on to generate offence, I’m not sure that my start of the third period would like like Krueger’s. I’ve highlighted two shifts that seem awfully curious to me. The shift with the fourth line is just sort of bizarre, in that it comes at the cost of an expected top two shift before the six minute mark. If you look down the table from the fourth line shift that ends at 2:53, you can see that Krueger was able to get four more shifts in before the six minute mark, with one for the RNH line and two from the Gagner line. If he hadn’t given the fourth line a shift there, he could have squeezed in an extra shift for the RNH line before the anticipated TV timeout at the first whistle after six minutes.

    I can at least see an argument for not shutting out the fourth line at ES entirely at that point of the game. The Oilers use Anton Lander to kill penalties and Teemu Hartikainen on the power play, they haven’t been on the ice yet in the third period, and maybe Krueger doesn’t want them sitting there with their legs getting heavy. It’s arguable and my general policy is that if I can see a case for what a coach is doing, I don’t get too crazy about it.

    The part that I find less defensible is coming back with Lander/Belanger/Hartikainen at the six minute mark. That’s a chance there to run four scoring line shifts in a row – two before and two after the TV timeout. It seems odd to me that, down a goal with fourteen minutes left, you wouldn’t take advantage of that but when instead give Lander/Belanger/Hartikainen a minute of ice time. This is the adjusting the sails of hockey coaching. Instead, the TV timeout happens 3-7 seconds after the RNH line members take the ice and the Oilers don’t really use it effectively. Worse, the Oilers’ take a penalty shortly after the game starts again.

    No objections to how they played the penalty kill – they basically held Gagner out of it, which makes sense. He and Hemsky only went out at the end, to try and quickly turn the PK shift into a scoring line ES shift, which didn’t work when they couldn’t get out of their own end of the ice. The Oilers were at least kind of able to use the TV time out effectively, with the RNH line getting a 28 second shift before it started and a 43 second shift immediately afterwards.

    The Oilers did use the TV timeout effectively here, they were just a little cramped for time because of the PK. They went Gagner/RNH/RNH/Gagner but just didn’t quite get the shift length out of it that they would have liked due to the penalty kill. Interesting that Gagner went out to kill a penalty this time after Lander took his spot on the previous penalty kill – I note that there was a faceoff and it may be that Krueger preferred Gagner for the faceoff to Lander. Kruger ran the same trick with Eberle and RNH that he’d run with Hemsky and Gagner on the previous PK and had them finish the penalty kill. It worked a little better this time and they got a shift out of it, which Krueger followed with the Gagner line. That shift ended relatively quickly with a penalty against Colorado and a TV timeout.

    That’s basically the hockey game. From that point, it was special teams and then Colorado iced things with an empty netter. Outside of the the first TV timeout in the third, it’s hard to find a lot of opportunities for Krueger to up the TOI of the Gagner line. I suspect that basically, you can’t do much better than 1/3 of the ES TOI in the crunch part of the game except in how the TV timeouts are used and the coach has to pick the group of people who are going to benefit from that. You can spread it around but there isn’t much point in doing so if you think that one of your lines is your best ES TOI line. There were some things working against the Oilers in terms of upping the TOI for the Gagner line there and, with the exception of the first TV timeout in the third, it’s hard to quibble with how Krueger ran his bench.

    One other Krueger point: he again pulled his goalie very aggressively when the Oilers had a PP late in the game. Dubnyk came out of the net with about 2:30 left and fifty seconds left in a PP. There’s reason to think that the Oilers are investigating this stuff from an analytic point of view; if anything, I’d like to see Krueger be more aggressive with pulling goalies.

    Email Tyler Dellow at tyler@mc79hockey.com


    6 Responses to Oilers’ ES Forward Deployment Down 2-1 To Colorado

    1. Doogie2K
      February 4, 2013 at

      Liked the point at the end about the aggressive goalie pulls. It’s always driven me nuts that NHL coaches are scared to death of pulling the goalie with more than about 75 seconds to go unless it’s a power play (when it goes out to around 90 seconds). It’s just not enough time to actually take advantage of the extra manpower – you get one, maybe two cracks at decent zone pressure before time’s up. This may be my cowbell; I dunno.

      Anyway, I’d like to see more coaches go this route and try the pull with around 2:00 to go, maybe more depending on manpower situation; at least that mimics the length of a standard power play. I see that a lot more in junior hockey, and it seems reasonably effective, at least by eye; I’ve actually seen very few ENG in that situation. Heck, one time I saw the Oil Kings coach pull his goalie with about five minutes to go down by three in Calgary; they didn’t score, but they didn’t get scored on, either, and they had pressure in the Hitmen zone for probably 70%-80% of that time. (And this was around 2008 or 2009, when the Hitmen were elite and the Oil Kings were kinda bad.) Extreme example, but it illustrates the point well.

      • Darrell
        February 4, 2013 at

        Studies using the existing pulled goalie scoring rates suggest that you should pull your goalie when down by a single goal with around 2:00 to 2:30 remaining. However, I do believe that there are practical realities (line changes, player usage, etc) that make this somewhat optimistic, or at least require careful analysis. What isn’t really in doubt is that pulling with 1 minute left is too late, and that you really should have the goalie pulled with 1:30 left.

        • dawgbone
          February 4, 2013 at

          I try to get my current team to pull the goalie with 2 minutes left I constantly get out voted and we do it with a minute left.

          Without fail, we come off and the guys say “man, if only we had another 30 seconds, we could have scored!”

          No shit.

    2. RiversQ
      February 4, 2013 at

      I didn’t check the situational breakdown, but 93 and 14 are playing 4-6min per game more than Hemsky and Gagner. That doesn’t seem justified to me. Hemsky at just 16min a game is underutilized.

    3. Elliotte Friedman
      February 4, 2013 at

      Interesting post, Tyler.

      Not sure if it still applies, but when I used to cover the NBA, those coaches always took out star players a minute or two before TV timeouts to get the extra rest.

      There were two in the first and third quarters (first whistle after 6:59 and 2:59) and an extra one in the second and fourth (9:59). That extra one rarely figured in, but the first two sure did.


    4. Jay
      February 5, 2013 at

      This is a wonderfully-written post.

    Leave a Reply

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