• Canucks/Kings line matching

    by  • April 16, 2012 • Hockey • 12 Comments

    Imagine a coach who made no effort whatsoever to match lines. He just hurled guys over the boards regardless of who the other team put out, although he did make some effort to get his better players on the ice more than his lesser players. We’ll call him Pat Quinn. Now, imagine his team just played the same team over and over. The other team wasn’t making any effort to match lines either; they were just tossing their guys over the boards but, again, trying to play their better players more than their lesser players.

    Over time, we’d expect that each team’s players would have time against the other team’s players that’s roughly proportionate to the time that those players played. If Quinn’s counterpart split the ice time 35%/30%/25%/10%, we’d expect each player on Quinn’s team to have about that much time against each line on the other team.

    Of course, that’s not how it works. Coaches (most of them) try to get certain matchups and work to do it. One of the cool things about the playoffs and the data explosion in hockey is that the pile of head to head games with coaches taking turns with last change means that we can see what a given coach is looking for and what he’s trying to avoid as far as matchups go by looking at how a players’ percentages match up against what we’d expect if nobody was matching any lines.

    Let me run through an example to show what I’m talking about here. In the two games LA played in Vancouver, Colin Fraser played 15.9 minutes of 5v5 time. Sammy Pahlsson played 19.5 minutes of 5v5 time. There were 80 minutes of 5v5 time in total. So Pahlsson played about 24.4% of Vancouver’s 5v5 time. If nobody’s fussed about who plays who, we expect Fraser to find that about 24.4% of his 5v5 time is against Sammy Pahlsson. In fact, only 18.2% of his 5v5 time is against Pahlsson. It seems that (barring randomness causing that) one or both coaches has decided that this is a matchup that they don’t want. You can create a sort of relative value here, by dividing 18.2% by 24.4% – you end up with 74.8% – that’s the percentage of time Fraser played Pahlsson relative to what he would if he just played equally against everyone.

    But wait! In G3 in Los Angeles, Pahlsson plays 24.3% of the 5v5 minutes. Fraser plays about the same percentage of time as he did in Vancouver – 13.9%. This time though, he plays 35.2% of his minutes against Pahlsson. Do our relative value math and you end up with 144.7%. Canucks’ checker extraordinaire Pahlsson spends a more significant part of his night shutting down Colin Fraser, which doesn’t seem like a particularly good use of Canuck resources. But then, Darryl Sutter has more say in these decisions now than Alain Vigneault.

    I’ve put together a table, from the perspective of the Kings, showing how their relative time on ice against various Canucks changed in G3 versus the first two games in Vancouver. Only players who played at least one game in each city are included. A positive number means that a player’s share of ice time against that player went up in LA as opposed to Vancouver and a negative number means that it went down. So, for example, Kings shutdown defenceman Matt Greene played a lesser share of his minutes against Burrows, Kesler and Sedin in LA than he did in Vancouver. Huh. Maybe Sutter didn’t know he had last change?

    The Kings lineup in G3 was basically as follows:

    Kopitar-Brown-Williams
    Richards-Carter-King/Penner
    Stoll-Lewis-Penner/King
    Fraser-Nolan-Loktionov

    Scuderi-Doughty
    Mitchell-Voynov
    Greene-Martinez

    How did things change? Kopitar-Brown-Williams saw a big reduction in their time against Dan Hamhuis and, to a lesser extent, Kevin Bieksa. They also saw substantially less of Booth-Kesler-Higgins. In exchange, they got more time against Burrows-Sedin-Hansen and and Lapierre/Raymond. You can note as well that they also saw reductions against Pahlsson and Malholtra. It seems fairly reasonable to think that Sutter likes that matchup with a wounded Burrows/Sedin line and likes being able to duck Pahlsson and Malholtra to an extent.

    The Doughty/Scuderi pairing looks to have been pretty aggressively matched with Booth-Kesler-Higgins in G3, relative to the first two games. You can’t tell from this chart, but they were each in the 50-65% range against those guys in LA, after being in the 35%-40% range in the games that were played in Vancouver.

    Richards/Carter were more of a pairing, so I’ll treat them separately. They played more against Hamhuis/Bieksa than in Vancouver, although they saw their ice time plummet against Edler/Salo – massive difference there. Sutter was also able to get them big jumps in ice time against the Tanev/Ballard pairing. (I should note – it’s a mistake to think of the road coaching staff as completely passive in these things. It’s more like a baseball draft where you get to take turns picking matchups, with the home guy maybe getting a few more picks early on.) They were also able to get away from Pahlsson. Funny thing though – they had a bad night chance wise. You play the cards but sometimes you crap out, I guess.

    The Mitchell/Voynov pairing, perhaps unsurprisingly, saw a big drop in their share of the ice time against the Kesler line, with a corresponding increase against Sedin/Burrows/Hansen. I’ve mentioned this before but there are really only so many different ways you can go and so many ways to set things up.

    I mentioned Pahlsson up above, so I’ll mention him again. The guys who saw the biggest increases against him were Greene, Lewis, Fraser, Stoll and Nolan, in no particular order. They were all much more tightly matched with him. I don’t think I’m going out on a limb here suggesting that Sutter had a preference for a power versus power matchup, running the scrubs against the scrubs, in a way that Vigneault didn’t do in Vancouver.

    In any event, I tend to think that this is pretty interesting stuff. It’d be awfully cool if TV put some resources into looking into this stuff. Watching the Canucks-Kings game last night, Bruce Boudreau sat silently for a while as Ron MacLean tried to make PJ Stock understand that assaulting someone after the fact doesn’t provide any protection to your star who’s just gotten whacked. Surely CBC has some interns who can spend time on timeonice.com during play and figure this stuff out – it’d be fascinating to listen to a coach like Boudreau talk about why he thinks LA might prefer to match Doughty/Scuderi with the Kesler line or why running power versus power is something that Sutter seems to have more of a taste for than Vigneault.

    (Oh – this stuff is pretty fun. It’d also be nice if the Oilers made the playoffs, so I could do this for a team I’m interested in.)

    About

    12 Responses to Canucks/Kings line matching

    1. Passive Voice
      April 16, 2012 at

      Wait, why would anyone ever opt for Greene against Sedin or Kesler? Especially if you had the option of Doughty/Scuderi or Mitchell/anybody?

      • Eetu Huisman
        April 17, 2012 at

        Because he’s a shut-down d-man many said he’d never be. (Tyler, maybe you should provide your readers a link when you’re referring something :-)

        • April 17, 2012 at

          See, that’s just a typo in David’s tweet. What he meant was, THEY WERE WRONG.

        • May 7, 2014 at

          Well good on you Ben for making the most of your chcane to see the Canucks. They sure played well against the Sens as they did against the Flames. I just wish they wouldn’t go to sleep when they have a lead and try to defend it. I’d like to see them offensively all the time, lead or otherwise. However, that’s just me. It’s also probably why Alain V. has relegated me to sharpening skates instead of suiting up for the fifth line.great post Ben. The game against the Redwings should be good. They are the team I enjoy watching most (after the Canucks) for the way they control the puck. Have fun.Dad

    2. Eetu Huisman
      April 17, 2012 at

      Couldn’t you do similar analysis for the regular season “series” of the Oilers against, say, the Flames? It would be interesting. (Or maybe someone should just convince Gabe Desjardins to add home/road splits of QUALCOMP to behindthenet.ca.)

      • dawgbone
        April 17, 2012 at

        You could, but there are different factors in play.

        For instance, the lineups for all 3 Canuck-Kings games have been virtually identical (barring line switches). Not only that, but both teams have the same schedule. You aren’t worried about one team playing back to back and their 3rd in 4 nights while the other team has their first game after a 3 day break.

    3. Arthur da Silva
      April 17, 2012 at

      Great analysis. Just a quick suggestion if you’re doing this in the future, could you colour code the table to make it easier to read?

    4. April 17, 2012 at

      Thanks for this.

      I’ve been doing a similar game by game analysis of this series over at Jewels from the Crown.

      Check it out if you’d like: here.

    5. Bulging Twine
      April 18, 2012 at

      Nice stuff man!
      Any idea what kind of a line matcher Daryl’s little brother Brent is? Seeing there is much speculation that he may be the next Oiler HC.

    6. April 23, 2012 at

      Once again the Canucks prove why they will NEVER win the Stanley Cup!!!! They don’t have what it takes and they NEVER will!!!! Cody Hodgson for Zack Kassian! Good grief who pulled that one off!!! Hit Henrik, hit Daniel, what else do they have? Quit sniveling, quit diving, and play the game!

      • Locky
        April 30, 2012 at

        I’m sure you appreciate the irony of rubbishing a trade made possible by line matching and quality of competition manipulation, in a blog post on the very same…

    7. David Staples
      April 26, 2012 at

      Roger Neilson said a coach can at best get one line match up that he wants each game, otherwise you interrupt flow of game and team too much. Even getting that one line match up can interrupt flow of team too much, so Neilson said he’d drop the effort in some games.

    Leave a Reply

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