• Matching Lines

    by  • November 9, 2011 • NHL • 18 Comments

    I’m finding the line matching this season to be particularly fascinating. This is the shift chart from the Oilers-Kings (aside: now that I follow European sport, I’m finding it increasingly difficult to remember which convention applies when listing – home team first or second?) game the other night:

    linematching

    It’s pretty obvious that Terry Murray was basically happy to play the line matching Tom Renney’s way. Renney was rolling out the RNH/Hall/Eberle trio basically as soon as Kopitar left the ice, leaving Horcoff/Smyth/Jones to play against Kopitar. With the last line change, Murray could have ducked that match-up but he didn’t seem too bothered by it.

    Up until last night, since the introduction of lines by seniority (Kid Line, Veteran Line), Renney has pretty much been able to get the Horcoff line a pretty good chunk of ice time against the other team’s better lines, even on the road. The Oilers have had five road games with the 94/10 pairing and the Kid Line so far. In Phoenix, L.A. and Calgary, Renney was able to run the Horcoff line against the other team’s presumptive best line. They went 9-8 in chances in those games (I’m just using Horcoff’s numbers for the line as a whole.) The RNH line went 10-8 in those games, with the numbers dragged down by a Hall-less game in Calgary.

    The Colorado and Montreal games though…yikes. Both of those teams ran what looks to be a third line against the Horcoff line, which went 11-7 in chances. The RNH line got attacked by the other team’s better lines and went 3-9 in scoring chances.

    A point that I don’t think we make often enough is that if both coaches seem to be pursuing the same match-up, somebody’s made a bad decision. Hockey’s a zero sum game and what’s good for you in terms of match-ups, by definition, has to be bad for your opponent. While it’s not many games, given what Renney’s done at home and the way in which the road games have gone – the Colorado and Montreal games were not good for the Oilers (four points notwithstanding), I think Renney’s got a pretty good handle on what gives his team the best chance of winning. Jacques Martin and Joe Sacco look to have figured it out too. I’m surprised at how Tippett ran his bench against the Oilers – he’s got four pretty competent lines and it would have seemed sensible to do the same thing.

    This is going to be worth paying attention to going forward, I think. I’m not at all sold by the Oilers’ start (sorry, guy who tweeted “Luck is a skill” at me last night) and I figure there’s some reversion coming. If there’s something to this (sample size etc.) and other coaches catch on, it won’t help.

    About

    18 Responses to Matching Lines

    1. Doogie2K
      November 9, 2011 at

      It’s a good thing I’m wearing a hat right now, because I think my head would’ve exploded at, “luck is a skill.”

      However you feel about “luck” in advanced stats, semantically or conceptually, that’s objectively ridiculous.

    2. November 9, 2011 at

      I’m largely in agreement with this post except that I’m struggling to see how Ryan O’Reilly’s line is viewed as a better trio than the Stastny/Duchene groups. For instance, as an NHL’er given the choice between Stastny-Hejduk and O’Reilly-Landeskog I’d much rather play the latter than the former.

      Which merely compounds the situation: why are NHL coaches (Martin excepted) in the early going willing to allow Renney to control the match-ups on the road? From observation and reputation, Tippett is a highly competent fellow and while I’m less wild about Sutter he did last two years in New Jersey and Lamoriello has no problem canning a coach he doesn’t like.

    3. Tyler Dellow
      November 9, 2011 at

      Jonathan -

      I just went through and looked at who played more. That decision was basically a toss-up – could go either way.

    4. Tyler Dellow
      November 9, 2011 at

      You can probably debate how to classify the Calgary and Colorado games, as far as whether Renney was dictating matchups. In the PHX and LA games, he seems to me to have obviously had what he wanted, with the other coach being fine with that as well. My concern is that the MTL game becomes the template for coaches going against the Oilers on the road – throw the third line out against Horcoff and focus on getting your two best lines against the kids and Belanger line. It seems like a pretty obvious play to me.

    5. Katie O'D
      November 9, 2011 at

      How do you think this will play out against the Bruins tomorrow? At the moment our best defensive forward (Bergeron) is playing with our best offensive forward (Seguin). I imagine Julien will try to get them out against the Kid line, since the Krejci/Lucic/Horton line has been susceptible defensively. Our third line is pretty good on de, so maybe Julien will do what you say the Avs and Habs did and put them against Horcoff? The home ice advantage will play a huge role tomorrow.

      • November 9, 2011 at

        My understanding from many Habs/Bruins confrontations is that Julien uses Bergeron the way Martin uses Plekanec, as a hard match. Last night, the habs 3rd line was pretty feeble and they still don’t have a 4th line so I expect Boston to be a tougher draw for the Oilers. The B’s 4th line is ok, if they still run Peverley/Kelly/whomever as a 3rd line, that’s pretty darn good too and then there is the matter of Bergeron and Krejci.

        My guess is Julien goes with Bergeron on the kids, Kelly on Horcoff. The only danger for the Bruins is if Horcoff can play significant minutes against Krejci, but Boston has more than enough depth to avoid this, especially at home.

        Also, I doubt the Oilers 4th line is good enough to break the opposition’s matchups on ennemy ice. Renney tried to send his guys out of sequence toward the end of the first and EDM’ 4th went up against Plekanec, who immediately plugged them with a SC and some nice offensive zone pressure. So there.

        The kids against Bergeron (who is basically the same kind of player as Plekanec, only much better) isn’t fair, but that’s life.

        • Tyler Dellow
          November 10, 2011 at

          What Olivier’s saying sounds reasonable to me. The kids have to make their bones.

          • Katie O'D
            November 10, 2011 at

            Yeah, sounds about right. However, Peverley isn’t 100% right now and Caron is still green. Also, with Paille out (and Pouliot on that line) the Bruins 4th line won’t be as strong in their own end. Horcoff’s line will likely cause some trouble no matter who they’re up against.
            Another thing to look out for: some of our de (McQuaid and Boychuk) have a hell of a time keeping their coverage in front of the net. I don’t see how this can end well against Ryan Smyth.

    6. E
      November 9, 2011 at

      yes, we were not impressed with your larvae. perhaps when they have pupated they will be more effective.

      could be that when both coaches go for the same match-ups they’re just working under different hypotheses about what skill set is more advantageous in the situation. the oilers, right now, seem to be at the far extreme of obvious matching. an assiduous sheltering of one line at home basically screams THIS IS OUR WEAK POINT HIT US HERE PLEASE PLEASE PLEASE PRETTY PLEASE to any coach with eyes to see, so i’m not sure why anyone would ignore that opportunity, but i can conceive plenty of other situations where the advantages/disadvantages of a particular match would be unclear or so nearly balanced as to make no certain difference. i doubt there’s always a ‘right’ and a ‘wrong’ to every match up.

      what i wonder about is why some coaches go through phases of not matching at all.

      • dawgbone
        November 9, 2011 at

        When both coaches like a matchup (i.e. Thornton vs Horcoff in an Oilers-Sharks game), I don’t necessarily think both are making a bad decision.

        For instance, San Jose is thinking that if Joe’s like can win the battle vs Horcoff, the Sharks should have a pretty good shot at a W. Likewise, Edmonton has to be thinking that if Horcoff can be about even against Joe, the Oilers have a better chance.

        • Tyler Dellow
          November 10, 2011 at

          Yeah but one of those coaches is more likely to be right about whether or not their guy is likely to win the matchup. If SJ is right, the Oilers aren’t organizing themselves the right way (assuming there’s a better way available) or vice versa.

          • Lobanovskyi
            November 10, 2011 at

            You’re right to say that in any given situation, one coach is more likely to be right (except where the likelihood of success is evenly split). However, without a sufficient number of instances from which to measure the likelihood of success of a line against another, you can’t tell which of the coaches is more likely to be right about who will win the matchup. One game is not enough. Eight games are not enough. Lines and D pairings change all the time. Seems to me like you have a sample size problem.
            I’m with E on this one, and agree that if a coach doesn’t match as a matter of course it seems like he’s given up.

      • Tyler Dellow
        November 10, 2011 at

        they’re just working under different hypotheses about what skill set is more advantageous in the situation.

        Yes, but one of them is right and one isn’t.

        i can conceive plenty of other situations where the advantages/disadvantages of a particular match would be unclear or so nearly balanced as to make no certain difference.

        Just because the coaches can’t tell doesn’t mean that there isn’t a right/wrong.

        I suppose you could say that it’s possible to have teams that are such that matchups make no difference. For example, a team with 20 equally skilled players facing another such team would be a case of this.

        • E
          November 10, 2011 at

          there might be a deep, underlying right/wrong, but given the aforementioned sample size problem, and the fluctuations in every player’s performance throughout a season based on injury/illness/fatigue/whatever, and inevitable need to shuffle lines, i think it’d be hard to derive a certain knowledge from past matchups that could be transformed into clear policies used to guide future matchups. plus, i would suspect that the percentage advantage on a particular match for the coach who is ‘right’ is so small that, considering how little time one line plays in a game and how seldom the same two teams will play each other with the same two lines, it’ll be totally drowned out by luck. really, the trick is to be winning on all four matchups, which is largely a function of how good your assets are.

          right now the oilers make an interesting case study for this kind of thing, but i don’t think it’s a generally applicable one.

        • Tom Benjamin
          November 10, 2011 at

          Other factors besides matchups matter, too. If you asked Alain Vigneault, he would say something like, “We worked at matchups three or four years ago, but now we figure if we play our game we will win regardless of matchups. We want the Sedins to take as many offensive end faceoffs as they can manage. We want Kesler or Malhotra taking defensive zone ones. With the options we have, I’d rather focus on getting my offensive players out in offensive situations than working to get the best matchup.”

          “I don’t want to let Tom Renney decide when I put out the Sedins and that’s where I end up if I try to get them out there against Nugent-Hopkins.”

          In other words, the best option for Renney can be to match and the best option for Vigneault can be to ignore the matchups. Renney is going for the more defensive strategy. Vigneault is going for the strategy that puts the Sedins in the offensive end the most. He’d prefer to see the Sedins out against someone besides Horcoff, but he’s not willing to give up any offensive opportunities to avoid it.

    7. Tyler Dellow
      November 10, 2011 at

      I just want to point out: two girls posted in this thread, a record for a hockey stats blog.

    8. RiversQ
      November 10, 2011 at

      Yeah I disagree about right/wrong with regard to certain matchups.

      It’s more about risk IMO. The safest bet is for both teams to go PVP for the true first lines, then the gentlemen’s agreement on the fourth line, and the rest usually falls in line. I don’t see either coach being wrong there regardless of the outcome. It will likely be the closest true talent matchup between the two teams. It’s just depth and the better players that rule the day there.

      Any other matchup breakdown means you will have significant mismatches that are not depth related. In those circumstances, the coaches are gambling and someone will be wrong. IMO.

    9. LazarusLongnap
      June 24, 2012 at

      With Yak now in the fold, and still no defense, our line-up conundrum has boosted for orbit. I was thinking: With this many horses, is line matching even worth the bother? I see there’s some good commentary here, but I’d to point out that your sound-bite game theory is a little too quick to the presumption of error.

      A point that I don’t think we make often enough is that if both coaches seem to be pursuing the same match-up, somebody’s made a bad decision. Hockey’s a zero sum game and what’s good for you in terms of match-ups, by definition, has to be bad for your opponent.

      There’s the difference between a lawyer and a mathematician: we don’t quit if the refutation of a cuddly home truth bores the jury.

      First, it’s not actually a zero sum game if your strategy today carries over into your competence on the ice tomorrow. Second, there might be hidden information, such as which player is nursing a cracked rib; this kind of thing might obviously parlay into a desire to evaluate a player in a specific role to inform upcoming roster moves. Finally, to be precise, a Nash equilibrium is a situation where–everyone having apprised themselves of the adversarial repertoire–no party can make a gain by changing their own strategy unilaterally. It happens quite a lot in a messy world that adversaries settle into a static engagement on rational grounds: you know you would do it different if you had all the information, but you don’t and the finding out is an ugly distraction.

      As one example, if the roster you face at game time is not the roster you practiced to play against, it could be counterproductive making late adjustments. A shrewd coach is not going to have kittens to slight advantage if doing so muddies up the peccadilloes. Hard to learn from your mistakes while dancing the deck chairs. In the big picture, clarity wins the war.

      As a second example, your wounded duck would be best shielded against the weakest opposing line, but if your moves to achieve this are correctly decoded, wounded duck will be on the menu. Better leave well enough alone. Perhaps your adversary suspects something from your passive response, but can’t narrow it down, and he already favours the situation in hand.

    Leave a Reply

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