• Na Zdraví

    by  • February 26, 2014 • Hockey • 5 Comments

    One of the positive things about getting interested in soccer for me is that it’s kind of changed how I think about hockey a bit. Soccer is a very low scoring game. As a result, it seems to me that there’s a bit more of a media focus on the process (everything leading up to a goal or a chance) than the outcome (goals) than there is in hockey. For whatever reason, hockey discussion doesn’t seem to include a lot of discussion about process – it’s all outcomes. Getting points? You’re doing good! Not getting points? You’re doing bad!

    A few years ago, the Oilers re-signed Ales Hemsky at the trade deadline. Hemsky was going through a pretty rough season as he came off a shoulder surgery but I was pretty thrilled with the deal. I’m not sure I’ve written anything that’s held up worse in terms of what I expected to happen (Hemsky returns to being a 70+ point guy per 82 games played) and what did happen (Hemsky becomes about a 40 point guy per 82 games played).

    My reasoning was basically twofold. One, based on what I’d read and heard, it seemed likely to me that Hemsky’s shoulder problems were behind him. I figured he’d probably be pretty healthy going forward, a guy who can be expected to play 70+ games a year. That part held up pretty well. Hemsky’s had two injuries of note over the past few years, both foot injuries, both caused by mistakes from his teammates. Jeff Petry made a bad pass in Detroit that a Red Wings defenceman fired into Hemsky’s foot, breaking it. Luke Gazdic shot a puck into Hemsky’s foot in Dallas, injuring it and causing him to miss some time. Other than one game he sat out after a hit, he’s otherwise been fine from a health perspective. Nothing chronic, nothing nagging that’s caused him to miss time or require surgery.

    The second part of my reasoning was that Hemsky had been an elite offensive player up through the 2010-11 season and that it seemed foolish to conclude that he was finished on the basis of one bad season in which he was coming off surgery. Fun fact: in the NHL’s dead puck era (1994-95 to present), Hemsky’s 28th in points/game (min. 300 games) from the age of 22-27 (for him, 2005-06 to 2010-11). That’s pretty elite offensive company.

    So my thinking was pretty straightforward: guy who piled up points at an impressive rate + no more nagging, chronic injuries = really good player to have around. I pulled a list of everyone who’s scored between 0.8 and 1.0 pts/gm from 22-27 (min. 300 GP) and then put together how they did from age 28-30 (2011-12 to 2013-14 for Hemsky).

    By some margin, Hemsky’s been the worst of the bunch. His decline is even worse when you consider where he started from – he’s produced points at 54% of the rate that he did from 22-27. Some of this is maybe attributable to a decrease in his TOI in general and PP TOI/role in particular but it’s a spectacular decline. He’s in a league of his own.

    It’s a weird thing, so weird that I don’t entirely know what to make of it. To my eye, he looks a lot like the same old Hemsky he’s always been but the NHL production has just disappeared. He led the Czechs in scoring in the 2012 World Championships and did it again at the 2014 Olympics despite some odd coaching decisions in respect of ice time. He produces very little in the NHL.

    Hemsky’s decline in boxcars isn’t entirely to do with a decline in his play – some of it has to do with ice time. I put together a table summarizing his age 22-27 seasons and 28-30 seasons to look at that.

    As you can see, the big change is in his PP TOI – he’s dropped almost two minutes a night as the Oilers pushed those minutes to their younger players. I haven’t always agreed with that – there are guys who, IMO, warrant PP TOI less than he does – but I can understand it. His ES TOI is actually pretty close to what it was during his salad days.

    A look at where the change in production has occurred is pretty telling.

    I suspect that the drop in PP production per 60 minutes is a bit misleading. There have been two changes that would, in my opinion, hurt his production. He doesn’t get much 5v3 TOI anymore and those are gravy minutes for a power play unit. He’s also playing on the second unit, which means generally not getting to start the power play (so pretty much always needing a zone entry to start) and playing with lesser players. If you do an apples to apples comparison, that difference would close significantly – the 5v4 numbers look to be a lot closer.

    That leaves the ES production. There’s no really good reason that I can think of for it to dry up. And yet, it’s fallen off 37% from Hemsky’s 2005-11 peak. It’s baffling. As I’ve mentioned a few times when it’s come up over the years, this is mostly related to a drop-off in the 5v5 shooting percentage when Hemsky’s on the ice. From 2007-11, it was 9.2%. Lots of goals, lots of points. Since then, it’s been 6.9%. His IPP has dropped off a bit too – from around 81% to 72%.

    We’re talking about reasonably large samples of shots with Hemsky on the ice here. The 2007-11 period is 1400 shots and the subsequent period is 979. I suspect that we can infer that Hemsky’s 5v5 on-ice shooting percentage was high in the 2005-07 period as well, which makes this decline even less likely. If you had a coin that said “SAVE” on one side and “GOAL” on the other that was weighted to come up goal 9.2% of the time, the odds of 68 GOALS in 979 coin flips are pretty low – around 1 in 150.

    A more likely explanation than “bad luck” for the decline in the Oilers’ 5v5 shooting percentage with Hemsky on the ice is that the shots over the past three years have been different somehow – less dangerous. The million dollar question – and I’m undervaluing this question here, because it’s actually worth more money to Hemsky and to teams that are considering trading for him or signing for him – is why this might be. What’s different when Hemsky’s on the ice the past three years versus 2005-11?

    I don’t have an answer for that. I can construct elaborate stories about how he enjoyed his greatest success with players who had no inclination to be the guy who carried the puck – guys like Shawn Horcoff, Ryan Smyth and Dustin Penner, who were happy to let him be the guy who had the puck and danced with it while they skated into space. The Oilers scored more goals per 60 at 5v5 with Hemsky/Horcoff on the ice than they did with Hemsky/Nugent-Hopkins. They scored more goals per 60 at 5v5 with Hemsky/Penner on the ice than they have with Hemsky/Hall. (Hemsky and Eberle have spent just over an hour together in their time with the Oilers due to them both being nominally right wingers. I confess that I’m kind of interested to see them together because Eberle’s probably the closest thing that the Oilers have right now to a guy who knows how to find goal scoring space in the offensive zone without the puck. He’s not superficially like Penner or Smyth, but he is in some other ways, if that makes any sense.)

    The question, though, is why that occurred. I don’t know the answer to that. Coming up with plausible sounding theories/stories is easy – proving them is a lot harder. Maybe the Oilers haven’t put the kinds of players around Hemsky with whom he’ll have success over the past few years. Maybe he’s into the Radek Dvorak phase of his career, being a really good third liner who doesn’t bring any offence.

    For whatever it’s worth, I took a look for players who had a lousy 5v5 on-ice shooting percentage through their age 28-30 seasons and rebounded. There doesn’t really seem to be a lot there. Amusingly, Penner jumps out. He’s a year older than Hemsky and his 28-30 seasons were spent mostly on the (offence) killing house floor in LA. The Ducks are north of 12% with him on the ice this year. Henrik Zetterberg’s age 28-30 seasons saw a pretty mediocre 7.6% on-ice shooting percentage; since then, he’s at 9.3%. Johan Franzen saw a similar leap. It does happen sometimes.

    My gut feeling is that he can still bring offence to a team but that he’s not a plug and play offensive player. He benefits from players around him who are comfortable playing offence without the puck, finding dangerous spots on the ice and creating space for him – if you go watch any Hemsky video on Youtube, the defining Hemsky move is hitting the blue line and slowing down, cutting towards the middle, while his linemates create new angles by skating. To me, that explanation squares more with what I see – he still zips around – and the fact that he still looks reasonably productive on the PP once you take the change in his role into account. I don’t know though – this is a blend of looking at data and gut feel.

    An emphasis on process when we talk about the game, like we see in soccer, in generating data that describes the processes in a hockey game, would shed some light on that. Hockey doesn’t have that, so you can go through the numbers you can find and come up with the most reasonable explanations you can given that and what you see. It doesn’t have to be quite this hard; someday, it won’t be.

    If I were a betting man, I’d bet that Hemsky’s remaining time as an Oiler is below ten days at this point. In 2012, I wanted them to bring him back because I figured his on-ice shooting percentage problems were variance that would wash away with two more years. Now, I’m somewhat less certain of that and much less certain that, if they did bring him back, they have the kind of people to play with him who could help change it.

    I’ve been a pretty unabashed fan of Hemsky during his time with the Oilers. He’s glorious to watch when he’s on his game and I have a natural sympathy for players who get on the wrong side of the media for no reason beyond “He doesn’t talk to us enough with a pleasant Boston accent.” Particularly when he gets unfairly tagged as soft or gutless when he’s a guy who’s taken hits for years. Plus, he’s the kind of player, you watch him do things and they’re the kinds of things you wish you could do as a player. Watching Hall go into overdrive is cool but anyone who’s played hockey knows what it’s like to skate fast, if not that fast. That can’t be said of sucking a defencemen into turning, putting the puck through his legs, doing that little hop over his trailing leg, going to the net and scoring with a trail of bodies in your wake.

    For his sake, it’s probably best that he gets traded. If the Oilers aren’t in a position to put players with him with whom he can succeed – (this started as a parenthetical then I realized it needed it’s own paragraph)…

    I don’t entirely say this as a criticism of them – I don’t think he’s easy to play with, although I also kind of think that they didn’t realize what they had when Hemsky/Penner/Horcoff beat the living hell out of their opposition when they played together from 2007-11, outshooting them 436-348 and outscoring them by a phenomenal 51-23. If you look at it on a year by year basis, you get this:

    The weirdly terrible numbers from 2010-11 are explained by the defencemen behind them.

    I’m not entirely sure that all of those people are real. Unsurprisingly, Horcoff/Penner/Hemsky were just fine in the limited time they spent on the ice with Smid/Gilbert, two Real NHL Defencemen, outshooting 16-11. The only other legitimate defenceman on that list is Jeff Petry. With two of Petry/Smid/Gilbert on the ice with those three, they outshot 21-15. That means that the rest of the time they played together in 2010-11 (75.98 at 5v5), a forward line that had crushed the opposition for three years got filled in to the tune of 33-58. I can’t believe I watched a single Oilers game that year.

    Where was I? Oh, right – if the Oilers aren’t in a position to put players with him with whom he can succeed, then I’m not sure why he’d want to be in Edmonton. It’s hard to blame the Oilers if they’ve decided that Hall, RNH, Eberle and Yakupov are the rock on which they’ll build their church. It’s not a criticism of him either – he’s made $40MM and you only get so many seasons in the NHL and scoring’s fun and winning’s fun. To quote the great philosopher Chris Pronger, it is what it is.

    If I was him, I’d want to find out if I still had being a 60+ point guy who can contribute to a really good team in me. If that’s the case, it’s hard to think of an offer from Edmonton that would make sense – it’s not going to be a life changing amount of money and, judging by his production, the fit hasn’t seemed to be there for three years. If I’m done as a point producer in the NHL, so be it – it happens to everyone eventually – but I wouldn’t want to be 38, sitting in a cafe beside the Vltava drinking a beer and wondering what would have happened if I’d left Edmonton.

    I’m not entirely unhappy that he seems very likely to be leaving – I’m very curious to see if someone, somewhere, can figure out how to make him produce offensively again and justify my third year of faith that he’ll turn it around. At the same time, he’s a historically significant Oiler, in the top ten all-time in assists, points and power play goals, who scored and set up some huge goals in the 2006 playoff push and run. I’m a little too young to comment on where he sits on the Oilers all-time list of exciting players to watch but it wouldn’t stun me if he’s a medalist. He’s been a real bright spot during some awfully dark years for the Oilers.

    Wherever he lands, within reason (sorry Vancouver), I’ll have my rooting interest for the playoffs. Here’s hoping it’s San Jose.

    Email Tyler Dellow at tyler@mc79hockey.com


    5 Responses to Na Zdraví

    1. ryan
      February 26, 2014 at

      sedins + luongo + hemsky and it still wouldn’t convince you? tough sell.

    2. Tyson
      February 26, 2014 at

      For some reason, I’m wishing you didn’t write this article. Perhaps I’m overestimating your influence over GMs in the league or underestimating their staff’s ability to see the same thing.

      ie. I don’t think illuminating the exact fit needed to have Hemsky produce helps his trade value.

    3. Caleb
      February 27, 2014 at

      Really cool post, obviously Hemsky is a fun an interesting case in Edmonton.

      I wonder though, is it possible to compare maybe using heat maps like Chris Boyle does in his work to see where Hemsky was taking shots or where shots were being taken when Hemsky was on the ice from 05-11 and compare that to his last three years in Edmonton? I don’t know if this has already been done or not, I’ve never seen it, but could that provide some light at all to help explain his on-ice sh%?

      Damn shot quality and shooting percentages.

      • Tyler Dellow
        February 27, 2014 at

        It’s not possible with the current state of data. I think you’ve asked a great question though and one that’s going to be to the future of hockey analytics. We’re going to get the motion data that the NBA has eventually. Once we have that, we’ll be able to start to ask questions very intelligently and get into things like how Hemsky’s play has changed. Maybe he pulls up less now. Maybe he tends to have fewer teammates driving the net.

        The sort of subtext of this post is that I think that there’s something going on that we’re all missing. Better data – which is coming – will allow teams to delve into those types of questions and explore how to maximize synergies.

    4. Scott
      March 9, 2014 at

      I think a lot of the stats-type bloggers and analysts dismiss shot quality as a real thing because there isn’t a way to quantify it. I’m glad to hear that you seem to add that just because there isn’t a way to quantify it yet, doesn’t mean it isn’t a real thing. As a fan who doesn’t have much statistical sophistication vis a vis hockey, I can’t believe that it isn’t a real thing. You know a good scoring chance when you see one. You sit upright. Your eyes widen. You know it.

      Teams that run a trap and counter-punch style of hockey do so for a reason: Namely, when the other team’s defenseman makes that bad decision exiting his zone you were waiting for, you get a breakaway or a two-on-one. Probably doesn’t lead to much offensive zone time. Probably doesn’t lead to much puck possession time, and your Corsi might suck, but you get a good scoring opportunity. You can get your Mike Cammalleri type forward in a good spot, even if your roster doesn’t have an Evgeny Malkin type player to skate around with the puck in the offensive zone for 45 seconds.

      In football, they measure big plays (typically any play of longer than 20 yards, but sometimes it’s defined as 40 or more). The best defenses concede very few big plays. You want to give your opponent as many opportunities to line up and make a mistake as you can. I don’t know what the equivalent statistic in hockey would be, but that turnover at the blue line that leads to an odd-man rush has got to be it. Heck, even in basketball, they measure fast-break points for and against. Why? Because those are easy scoring opportunities.

    Leave a Reply

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