• MacTchecking: Taylor Hall and Matchups

    by Tyler Dellow • January 13, 2014 • Hockey • 7 Comments

    Craig MacTavish did a lengthy interview with David Staples that Staples has published in four parts over at the Cult of Hockey. There’s a lot of interesting stuff in there that catches the eye and I might come back to some of it over the course of the week. Let’s start with this:

    I said this before, when I took over my one mandate was to turn this team over to the guys who are ultimately going to lead us out of this, and that was largely the draft choices, Taylor, Nugent-Hopkins, Eberle, and put them in very prominent positions. No hiding them in the match-ups. Turning them loose to the toughest match ups. And at the start of this year, that didn’t go well. It went poorly. But I believe strongly failure is a great motivator. I think that those failures really helped those players try and restructure their game and the way they think the game and some of the decision making that they make during a game.

    This was kind of surprising to me because I was sort of under the impression that this had been done last year and that Hall, Eberle and RNH had handled it reasonably well. Hall posted a 49.5% Corsi% with the score close last year; this year he’s down to 44.4%. It seems odd to me that we could explain this by reference to tougher competition.

    Just as an aside: I’m somewhat skeptical of the value of line matching in general. While there are some exceptions, there’s a problem. Hockey’s a zero sum game. If you run Taylor Hall against Sidney Crosby, you’re blunting Hall’s effectiveness but you should gain lower down the lineup from your lower lines not having to run up against him. Hall’s going to get his twenty minutes. Crosby will get his twenty minutes. However they’re deployed, there’s costs and benefits. Outside of the obvious – letting them feed on a hapless fourth line or third pairing – I’m skeptical that you can really create a ton of value by how you match your lines. Someone’s got to play the minutes.

    On that note, one of the interesting things as I go through this is seeing how the defencemen against whom Hall is matched change from games played in Edmonton to games played in other cities. Basically, if it’s a road game, no third pairing time for Hall. If it’s a home game, he tends to get some third pairing time and more second pairing time. If nothing else, this a good exercise to get a sense of the matching that’s going on and how it looks in a couple of different spots.

    Anyway, MacT’s statements imply that the Oilers were hiding Hall with matchups, his mandate was to get them going against the best and that that was part of the reason for the early season struggles. Let’s take a look at Hall’s matchups this year. I’m going to focus on teams that the OIlers played home and away last year that they’ve played home and away this year in order to have decent comparators.

    San Jose

    The Oilers have played the Sharks twice, once in Edmonton and once in San Jose. Neither game was all that competitive. The Sharks came to Edmonton and were up 3-0 by the 23:45 mark of the game. The Oilers went to San Jose and were down 4-0 by the 29 minute mark. Still, a couple of interesting things jump out.

    First, I wonder if Eakins was chasing a defensive pair with Hall in the Edmonton game. That’s a pretty incredible matchup of Hall with Braun and Stuart. It seems unlikely to be a coincidence. Hall’s matchups with the defensive pairs in San Jose was much less consistent – a much flatter spread.

    As far as the forwards go…look, I’m a guy who thinks Joe Thornton is still the guy on the Sharks. Thornton didn’t play much against Hall when these teams met in Edmonton. In San Jose, where Todd McLellan had last change? McLellan ran him at Hall all night long. When Hall wasn’t benched. If I was trying to guess the matchups that the coaches were looking for in these games, I’d guess Eakins wanted that defensive pair in Edmonton and McLellan wanted the Thornton line in San Jose.

    I wouldn’t characterize the Edmonton game as Hall & Co. being turned loose against the tough matchup although the score probably had something to do with this – if you’re trying to get back into a game, it probably makes sense to send your best offensive players after the most vulnerable defensive pairing that’s readily available. Coaches can hide one pairing on the road to a degree but you can ‘t hide two. That said, Eakins was after that matchup before the Sharks scored, so it wasn’t just an attempt to get back into the game.

    What about last year? Again, we see that McLellan had a pretty clear group to play against Hall in San Jose; again, it was Joe Thornton and Joe Pavelski, this time with Patrick Marleau. He had a defensive pair that was pretty much hard matched against those guys too – Stuart and Vlasic played close to 70% of Hall’s minutes in San Jose.

    If we look at the games in Edmonton, it’s harder to find a clear matchup like we saw earlier this year. Those line matches look pretty much like about what you’d get if you weren’t matching lines and Hall was just matched up against people according to how much they played. Of note, those minutes are probably still tougher than the ones he played this year in Edmonton, given that he didn’t really play against Thornton in Edmonton this year and he did in 2012-13.

    As to MacTavish’s point then…I don’t really see evidence of it here.

    Chicago

    The Oilers have been smoked by Chicago in Edmonton in a game that was over early. They lost in Chicago at the end of a road trip in a game that was pretty back and forth – it was tied after two periods Chicago didn’t really salt it away until the third period.

    What does the line matching show? Well, again, I think we can make a case that Eakins is hunting the second pair with Hall in Edmonton. It’s a different story in Chicago, where the Hawks had the Seabrook/Keith pairing out against Hall about two thirds of the time. That doesn’t really seem to fit with running Hall at the hardest match-ups.

    The forward matchups tell the same story. Smith/Kruger/Bollig played twice as much against Hall in Edmonton as they did in Chicago. Toews/Sharp played much more against Hall in Chicago than they did in Edmonton. Again, hard to find support for MacT’s assertion here – we can see what being matched up against Chicago’s best looks like in the game in Chicago.

    Last year, Hall got a lot more Duncan Keith in Edmonton, along with getting the heavy Duncan Keith matchup in Chicago. He also seemed to generally get tougher minutes as far as forwards in Edmonton than he did this year – Marian Hossa and Patrick Sharp were in his three most common opponents in Edmonton, not so this year. Not much Toews at home, lots of Toews on the road.

    This data doesn’t include tonight’s game, but I kept my eye on the matchups, given that I most of this post in the can. Joel Quenneville ran Toews/Hossa/Sharp at Hall but, for whatever reason, he didn’t match Keith/Seabrook against them. Instead Hall got Hjallmarsson/Oduya. Keith/Seabrook got a pretty heavy does of Hemsky. Interestingly, to me anyway, Hemsky and Hall were together the last time the Oilers came through Chicago. Maybe Quenneville still has some time Hemsky.

    Columbus

    There’s not really a lot to see here this year. The Blue Jackets game in Edmonton turned into a blow out pretty quickly. Eakins probably wasn’t too fussed about who his guys played against. Certainly the defensive pairings don’t really suggest much in the way of line matching – I find it funny that the Blue Jackets match Jack Johnson so heavily against Hall in Columbus, given that Jack Johnson. Eakins did seem to go after the Jackets’ top forwards with Hall – he played more than 2/3 of his time against Foligno, Johansen, Umberger, Jenner and Atkinson – but it was a weird game due to the blowout factor.

    Hall doesn’t really get after Columbus’ third defensive pairing as much when he’s away this year – again, this doesn’t suggest that Eakins is actively seeking out the toughest matchups at home.

    Last year was pretty similar for Taylor Hall. Tons of Jack Johnson in Columbus and then a smear of defencemen in Edmonton. There doesn’t really look to have been a lot of matching of forwards either – it kind of looks like Krueger wasn’t too worried about it.

    Calgary

    The defence match against Calgary has been pretty consistent home and away. Giordano only played one of the two games in Calgary, which is why his number is low for the Calgary games. Kind of bizarre that the Flames have played TJ Brodie so much against Hall in Calgary but it’s really a function of not having a lot of options, I think.

    Calgary doesn’t have much in the way of forwards this year, so it’s awfully tough to identify a tough matchup. That said, Hall did seem to play more against Calgary’s sketchier players in the Edmonton game than he has in the Calgary games. The Oilers were trying to come back in that game in Edmonton but Hall saw a lot more of Lance Bouma, David Jones and Joe Colborne than he has in Calgary. Again – hard to see Eakins chasing the tough matchups here despite having last change. I’m not critical of Eakins for this; if you’re chasing the game, you’re chasing the game, but it’s not the toughest matchup.

    Looking at last year, the pattern looks similar. The defensive matches in Calgary saw a hard match with Giordano, who appeared in all of the games. Bouwmeester only appeared in one game against Edmonton, so that’s a pretty hard match too. In Edmonton, the time against the various defencemen seemed to be a little more evenly spread out.

    The forwards are pretty interesting. There’s a clear match that Bob Hartley liked in the Calgary games – lots of Glencross, Tanguay, Iginla and Stajan. Cammalleri’s TOI against Hall is really low, given that he played in two of the games. Things were a lot different in Edmonton, as Hall was kind of sent after the soft underbelly of the Flames team. Sort of like what happened this year.

    Colorado

    There’s a very clean match for Hall in terms of forwards, both in Colorado and in Edmonton: he was matched up against Parenteau/Stastny/Landeskog in both cities. Harder to see a defence matchup in Edmonton – as I recall, Hejda got hurt early in that game, which probably meant Colorado was scrambling things. Colorado was able to protect Guenin and Barrie in Colorado; not so in Edmonton.

    Has that changed from last year? Not really. The forwards are a little different but there was a pretty strong match up front for Hall last year in Colorado games in Duchene/Parenteau. The Avalanche had some injury/contract problems last year, which probably explains Landeskog being further down the list. As far as the defence go last season, we see the same thing we’ve seen with most of the teams that we’ve looked at: Colorado was able to shelter certain defencemen better at home than they were in Edmonton. Tyson Barrie saw a lot more of Hall in Edmonton than he did in Colorado.

    You can kind of make an argument that Hall’s getting the tough matchup here but it’s no different than what he was getting last year. It doesn’t seem to me to be something that’s changed this season.

    Dallas

    Oh look: another home game in which the Oilers spent most of the night trailing and chasing the game. Dallas scored about ten minutes into the game and then held on the rest of the way, adding a pair of empty net goals late. Again, you’d be hard pressed to argue that Hall saw the toughest matchups in that game. He barely saw the ice against Dallas’ most dangerous offensive players, unless you really rate Vern Fiddler.

    Notably, Hall looked to be not facing Benn/Nichushkin/Seguin very much in Edmonton. He saw lots of Benn in Dallas, although not very much of Nichushkin. Seguin was ill and didn’t play.

    There’s a real mix of defencemen in there, although Dillon and Robidas saw most of the minutes in the home game. It’s a pretty stark contrast with the road game – Robidas was hurt by the point and I would guess that Lindy Ruff has maybe decided that Taylor Hall is something other than a player to play behind Matt Read at the World Championships. Both Alex Goligoski and Trevor Daley were up over 80% in terms of the percentage of Hall’s minutes that they played. He certainly looks to have had a matchup that he wanted.

    Looking at last year, Hall looks to me to have been more cleanly matched up against Dallas’ best forwards – it wasn’t a hard match or anything but the guys who played most on Dallas are the guys against whom Hall played the most. It’s not a third line scenario, like it was this year. The defencemen are a bit of a smear – Hall saw pretty significant time against everyone but Jamie Oleksiak.

    In Dallas, there was a pretty clean matchup against Derek Roy, Antoine Roussel and Erik Cole. The Dillon/Robidas pair saw most of the defence minutes, although Glen Gulutzan didn’t seem too fussed about any particular defensive match.

    Overall? The six F who Hall saw this year in Edmonton don’t really look to me like the hardest matchup available.

    Phoenix

    I’ve got a lot of time for Dave Tippett as a coach. I don’t know that there are a lot of coaches who are using data intelligently but Tippett’s impressed me as a guy who makes big improvements happen. Listening to him talk, particularly about the value of a guy who’s shakier in his own end but spends less time there, I kind of think that he thinks about hockey the right way. He seems to take a collection of spare parts every year and mold them into something competitive. He gets results, without much talent. Sure, the hockey’s horrible to watch but that’s not his problem.

    Starting with this year, there’s no particularly strong forward matchups in the Edmonton game. Keith Yandle’s number seems curiously low and Ekman-Larsson matches up against Hall in 54% of his minutes. Now slide over to what happened in Phoenix – do you think Tippett had a plan? Hanzal, Vrbata, Korpikoski, Boedker and Vermette eat up more than 2/3 of the ice time against Hall that was played by forwards. Mike Ribeiro probably couldn’t tell you if Hall played in the game.

    Now look at the defence number: Ekman-Larsson and Connor Murphy both play over 70% of Hall’s minutes. That’s a pretty amazing number. It’s interesting to me that Yandle’s number is so low – if you read the embedded pieces on the US Olympic Team, someone had a line about how opposing coaches would be thrilled if Yandle ended up in the American top four. If Dave Tippett is expending coaching bullets to avoid Yandle/Hall, it kind of looks like he agrees.

    Looking at last year, there’s a bit of a smear of TOI amongst the forwards in Edmonton, although Gordon/Moss/Klinkhammer seemed to get more minutes than you’d expect, given their overall TOI. Contrast that with the forward TOI in Phoenix. Vermette, Doan and Nick Johnson, whoever that might be, carried the load there.

    It’s the defence that are really interesting though – we see the same thing again. Big, big minutes for Ekman-Larsson, who got up over 80% in both Edmonton and Phoenix. Tiny numbers for Yandle. Hmm.

    Stepping back and looking at the big picture, for the games in which the OIlers have played home and away against teams that they played home and away last year, it’s hard to find any evidence of Hall getting some sort of a push into crazy minutes. Overall, there doesn’t seem to be a lot of change. At home, where Eakins has more control over the matchups, Hall seems to be getting matchups as easy or easier than last year.

    So it’s hard, looking at these games, to give a lot of credence to MacT’s statements about Hall’s ice time and who he’s matched up against. Last year, Hall was (by inference) doing a much better job at holding his own in the minutes that he played. This year he isn’t. Matchups don’t really seem like a plausible argument for that to me.

    One final note, more general than specific: line matching is a pretty fascinating subject about which little is known. I remember when I was a kid hearing debates on TV on the merits of matching versus rolling the lines. As I said above, I’m skeptical that there’s a huge advantage to be gained. One line of thought that does make sense to me – if I was a coach, I think I’d be more likely to go power versus power in games that I was winning. In a game like that, I’d rather reduce my chances of scoring and their chances of scoring by trying to get the big guns on the ice together to beat each other senseless.

    If I was behind, I think I’d go the other way: get a player like Hall on the ice against the opposition’s weaker players at every possible opportunity. If it means running the Gazdic line against Jonathan Toews for a shift and praying, so be it. The losing team should want to ramp up the possibility of the goals and hope that the hockey gods look upon them with favour; the winning team should want to reduce that as much as possible.

    All in all, it’s a very fertile area for research and one that I’d bet will be mined heavily over the next few years. I’m looking forward to someone really figuring out the theory and proving what works and what doesn’t.

    Email Tyler Dellow at mc79hockey@gmail.com

    About Tyler Dellow

    7 Responses to MacTchecking: Taylor Hall and Matchups

    1. KSC10032
      January 13, 2014 at

      Current trends seem to focus more on forward lines “power v. power”, sometimes looking for the edge against defense pairs more than forward lines. Its also more difficult when 4 lines are involved given the drop-off in quality involved with most teams’ 4th line personnel.

      As an oldtimer, I remember the hockey of the 60′s, 70′s, and early 80′s, where teams rally only played three lines and 4-5 defensemen every night (teams could only dress 19 players). The normal line structure then was to have a #1 line with the best available of each forward position always playing together, a 2nd line and a “checking” line. Virtually every team played their checkers against the other side’s first line and vice versa, with the 2nd lines playing each other. There seemed to be a gentleman’s agreement that this would be accepted without a lot of juggling. Coupled with (much) longer shift times, line-matching was a lot simpler.

    2. Zip
      January 13, 2014 at

      Hi Tyler, in reference to your closing paragraph; how might an aspiring number cruncher go about proving whether or not line matching is effective? What variables would have to be considered, and what would be necessary for results to be conclusive?

    3. tphillers
      January 13, 2014 at

      Excellent work, Dellow! I had my suspicions that MacT was blowing smoke all over the place because Hall, by my eye, has been deployed rather craftily by Eakins. Your data shows that there is something more complex going on with Hall’s TOI than simple power v. power. I wonder what RNH’s numbers would look like in comparison. It has seemed to me that RNH has been getting the bulk of the tough assignments relative to his talented teammates.

      On the other hand, just as a point of discussion, I think I might quibble with your line matching theory you close with. I am by no means an NHL coach but it would seem to me that offense is generated in a hockey game through sound, deliberate defensive play that pushes the puck out of the d-zone and toward a scoring area. This takes talented defensive players and puck-movers (i.e. Pronger) to slow down the play and reset the offensive schema. I recall how 44 would consistently skate back into the defensive zone, in possession of the puck, before moving it forward with the right personnel, in all situations, too (it was most aggravating with four minutes to go in a 2-goal game). That poise was as invaluable to generating offense as having the best offensive weapons on the ice. So, having said that, I am not sure I would want to willingly expose my team to a situation where Gazdic is facing Toews late in the 3rd period in a close game and praying for the fortune’s of the hockey Gods — I’d lose more than I would win with that matchup; rather, I think I would want to stay the course with my power v. power matchup, and make sure my best defensive pairing never misses a shift v. Toews, instead of sending Hall over the boards insistently (which makes statistical sense) at the expense of a balanced approach. I understand your philosophy, but I think line matching is a very delicate process that requires more attention to every area of the ice rather than simply deploying your best offensive weapon as frequently as possible when trailing in a close game, and leaving your weaker lines vulnerable to a moment when the opposing coach can put out his most talented line to seal a victory.

      As for researching these theories, I have no idea where to start but I assume marrying by-the-eye analysis with hard data would be an important methodological step. But that’s more your bread-and-butter than mine.

      Thanks again for the hard work.

      –tphillers

    4. WeirsBeard
      January 13, 2014 at

      Any serious differences in ZS year vs year? Is this the uptake in competition MacT is referring to?

      Makes sense that the opposition for Hall hasn’t really changed. The Oil personnel hasn’t really changed year over year, we haven’t suddenly lost or gained a new high end line nor an elite shut down line (a la Anaheim with Marchant et al).

      • godot10
        January 13, 2014 at

        Under Krueger, Hall faced the toughest opposition (which he has faced since year 2), with the best linemates mostly, and advantageous zone starts.

        Under Eakins, Hall has face the toughest opposition with less than best linemates a lot of the time, with less advantageous zone starts.

        i.e. Krueger wanted to give Hall a chance to win the game, while Eakins is trying to prove to everyone that he is a smart coach. A coach that is trying to prove to everyone he is a smart coach is a bad coach.

    5. Nick
      January 13, 2014 at

      Thanks – agree with your insight re: line matching and when it may or may not be useful. Am inclined to be a bit more lenient re: the scope of MacT’s comments. Perhaps Hall isn’t getting the toughest matchups game in game out but I imagine you’d be hard pressed to show that Hall/RNH/Eberle are being actively sheltered, even when they are getting slammed. His interview was really interesting and I’m greatly looking forward to the rest of your analysis.

    6. Showerhead
      January 16, 2014 at

      Late, but I just needed to share my enjoyment of “MacTchecking” as a pun. Use it again someday!

      This Hall business is worrisome. Someone mentioned on the twitter the other day that he skipped an optional skate to rest his shoulder, which was even more worrisome overall.

      I had a pet theory for a while that Hall was flying the zone / cheating for offense more this year in an attempt to get his pre-Olympic boxcars up but I don’t think the shots-for numbers support it.

    Leave a Reply

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