Skip to content

Awareness of analytics a necessity for new NHL executives

Apr 11, 2014, 11:35 AM EDT

Maple Leafs Panthers Hockey AP

In the past week, former players Trevor Linden and Brendan Shanahan have been tapped to lead two of the most valuable franchises in the NHL.

Plenty has been written on these moves, but it’s interesting that the topic of analytics has come up in both cases. Ten years ago, you can pretty much guarantee it wouldn’t have.

On Linden’s hiring in Vancouver, here’s part of what Nick Cotsonika wrote for Yahoo Sports:

Though Linden looked and sounded great at his press conference, he said little of substance. Though he said he believes he’s “ready for this challenge,” he was unconvincing when it came to why. Example: Linden told reporters on the side that he knew little about analytics, noting the trend arose after his playing days.

That doesn’t mean Linden will fail. That doesn’t inspire confidence, either.

On Shanahan’s hiring in Toronto, here’s part of what Michael Grange wrote for Sportsnet:

Why does the richest organization in hockey not have a single staff member devoted to the emerging field of hockey analytics when a bunch of numerate hockey hobbyists were predicting the Leafs’ demise for free, on Twitter, for months?

For that matter, why didn’t the richest organization in hockey invent hockey analytics?

Why don’t they already have 20 of the smartest, geekiest hockey fans in the world locked in a warehouse somewhere with a wall of computer equipment and video archives inventing cutting-edge ways to understand the game?

Hockey has been a relative late-comer to the revolution in sports analytics. Baseball was the trailblazer. It’s a big part of basketball now as well.

The so-called “advanced” hockey statistics — which basically just try to measure the amount of time a team controls the puck — aren’t perfect, but when the top four Fenwick teams are Los Angeles, Chicago, San Jose and Boston, and the bottom three are Edmonton, Toronto, and Buffalo, there’s probably something to them.

Of course, there are anomalies — New Jersey is the fifth-best Fenwick team, while Colorado is the fourth-worst — but that doesn’t mean the statistics are useless. To dismiss them just because everything doesn’t line up perfectly straight is foolish.

In Vancouver, Linden will need to be a fast learner on the topic, possibly with the help of assistant general manager Laurence Gilman, who survived the firing of GM Mike Gillis.

In Toronto, Shanahan may need to not only learn; he may need to help convince the organization that analytics can’t be ignored any longer.

  1. patthehockeyfan - Apr 11, 2014 at 11:38 AM

    The picture made sense before I read the post.

  2. thailer35 - Apr 11, 2014 at 11:52 AM

    Toronto, if you see this post, I would like an interview.

    • letsgolightning - Apr 11, 2014 at 12:01 PM

      You read an article mentioning hockey analytics. You’re overqualified.

  3. rdk - Apr 11, 2014 at 12:17 PM

    Masai Ujiri is the only person who knows what he’s doing over there.

  4. shoobiedoobin - Apr 11, 2014 at 12:20 PM

    No matter how complex the mathematical formula, it will never quantify hockey. This sport is too fast with too many wildly unpredictable variables and intangibles. Stats are for people who don’t have to chops to give a game said eyeball test, and the more someone wants to be able to and can’t, the more complex the formula will get. We all want to be experts and this is a byproduct of only a few actually truly understanding this game. The rest conjure up silly fenwick and corsi math. Don’t get me wrong-it accurately reflects exactly what the numbers say but those numbers don’t quantify hockey. Never will.

    • timmmah10 - Apr 11, 2014 at 12:33 PM

      ^^^
      Is this Brian Burke?!?

      • shoobiedoobin - Apr 11, 2014 at 12:45 PM

        If Brian Burke said that it’s the only thing he’s ever been right about.

    • imleftcoast - Apr 11, 2014 at 12:46 PM

      Said everyone who doesn’t understand analytics. They don’t replace anything – they supplement hockey knowledge. If every detail can be the difference, a team can’t afford to ignore analytics.

      • shoobiedoobin - Apr 11, 2014 at 12:48 PM

        They supplement nothing. You can toss all the corsi BS you want, someone with an eye for the game and hockey sense won’t need them.

        They do act as a guide but really it’s not much. The payoff isn’t there. You have to cook up this hilariously complex formula just to get a slight kinda-sorta helping hand? Not even worth it.

      • salmon90 - Apr 11, 2014 at 3:15 PM

        Complex formula?! – it’s actually really simple math.

        More shots directed at opposition net = better puck possession= better chance at sustaining long term success

        Less shots directed at opposition = less puck possession = worse chance at sustaining long term success

        Corsi and Fenwick are general guidelines for measuring which teams are better puck possession teams – who would argue that worse puck possession teams are set up better for future success?

    • grocerystick101 - Apr 11, 2014 at 1:23 PM

      No one is saying that analytics should replace conventional hockey knowledge. I don’t think anyone is proposing that NHL teams start hiring mathematicians from MIT who have never laced up a pair of skates as their GMs.

      It’s a supplement to conventional hockey knowledge, and in the case of the Leafs it is a clear example of advanced statistics correctly predicting a team’s demise. It may not be a perfect system, but it’s ignorant to ignore it completely.

      • shoobiedoobin - Apr 11, 2014 at 1:28 PM

        No one’s saying that? You speak for everyone? Funny, it’s all I ever see online. It’s fun to think that you’ve “cracked the code” and can see the future (because that’s the fantasy being satiated) but it’s possible the mundane is also accurate-that the corsi numbers were right but not because they knew they’d be, but just by chance. I’ve seen corsi predict team failture for teams who spontaneously, on a dime, turn it around. If you need this to supplement anything, I feel bad for how you think.

      • grocerystick101 - Apr 11, 2014 at 2:11 PM

        Of the top 16 teams based on Fenwick ratings, 12 are currently in playoff position. Of the bottom 14, 10 are not. Last year, those numbers were the exact same. That is too overwhelming to be dismissed as coincidence. Again, not something that should be relied upon exclusively and perhaps it is overemphasized by the media, but there is clearly come correlation between these ratings and on-ice success that you can’t ignore.

      • salmon90 - Apr 11, 2014 at 3:16 PM

        I’m guessing shoobledoobin’s team is one of the lower ranked puck possession teams around…

      • shoobiedoobin - Apr 11, 2014 at 7:34 PM

        So who is my team then Salmon? Let the great bias witch hunt commence. Don’t put in too much effort, you wouldn’t want your easy answer to stop being so easy.

        I love how it’s now officially boiled down to simple puck possession stats. Don’t tell that to me, tell it to your brethren net-wide who argue that it quantifies the game. They already had stats for that by the way, which suggested the same thing. The necessity for it to be more complex than previous, wasn’t necessary at all.

    • hsnepsts - Apr 11, 2014 at 7:20 PM

      You’re right, if you’re talking about the scoring column in the Saturday paper.

      That’s not the analytics that they are talking about here.

      Over time, on average, fundamental understanding paired with quantitative analysis is going to beat fundamental understanding alone.

      • shoobiedoobin - Apr 11, 2014 at 7:30 PM

        So it’s went from “this is an eyeball test substitute” to “this is valuable and needs to be a part of the game” to “hockey sense still prevails but you need analytics too”.

        Lordy lordy…when the stat freaks come to a group conclusion on the value of these things, send it to a vote and get it over with. I’d ask if you even read half of what I’ve been saying but I’ve dealt with you on wings forums before, you’re the confrontational closet-troll type.

  5. exsane - Apr 11, 2014 at 12:38 PM

    @shoobiedoobin

    You nailed it right on the head.

  6. dbfan4ever - Apr 11, 2014 at 12:47 PM

    We aren’t supposed to dismiss analytics because of a few anomalies but yet most of the experts would have us dismiss the anomalies like Colorado because they rank poorly on the analytics! Can’t have it both ways! Maybe when things like confidence, heart, team chemistry and performing in the clutch can be quantified I will start paying attention to analytics!

    • shoobiedoobin - Apr 11, 2014 at 12:56 PM

      Bingo. I could watch a guy for a couple weeks and tell you more about him than any stat ever could. Anyone who knows the game well enough could. The smartest hockey minds aren’t putting much stock into numbers, this isn’t baseball where that sort of thing might be possible. You think Scotty Bowman needs to stare at a corsi sheet that indicates 4th line dusters stack up against elite players in some contrived laughable way? You think Mike Babcock won a gold medal because he, Ken Holland and Steve Yzerman sat around gawking at fenwick figures, or did they use the hockey sense you naturally have?

      Like I said, those who defend these saying they’re more than they are, they do it for a reason.

  7. hsnepsts - Apr 11, 2014 at 12:47 PM

    I am both a big time hockey fan and a statistician.

    I work in the stock market, but my passion is hockey stats. I WISH I had access to a full database of game records.

    Years ago I had the idea to approach the Canucks, to see if they would want to take me on in some kind of full time or part time capacity. But I never did. I figured they probably already had guys like me hanging around. My mistake.

    • spezzdispenser - Apr 11, 2014 at 1:24 PM

      It’s never too late, child. Pick up the phone today.

      • hsnepsts - Apr 11, 2014 at 6:58 PM

        Anyone have Trevor’s number?

    • imleftcoast - Apr 11, 2014 at 2:12 PM

      You know stats and how to spell Snepsts. Harold playing for as long as he did would be an anomaly. No model would predict that.

  8. purpleguy - Apr 11, 2014 at 1:01 PM

    Gee, fans of the most old-school, traditional sport short of baseball criticizing a new wrinkle to assist analyzing a game, instead saying all it takes is a “hockey eye”? Shocking. Probably used dial-up internet to get to this site.

    • shoobiedoobin - Apr 11, 2014 at 1:10 PM

      That’s the exact assumption people like you would naturally jump to. There probably are people like that but the reality is some people are smart enough to know that these stats are what they are. It’s the people who want, not think, but WANT them to be more than they are who are wrong. Now why would a sports fan want a stat to solve the game for them? Think about it.

    • imleftcoast - Apr 11, 2014 at 2:21 PM

      Baseball was also a game that most said was not subject to stats and required a ‘baseball eye.’ There were so many intangibles: guts, desire etc. Over a broad spectrum of stats, the anomalies fall away, and you’re left with a model that can supplement hockey knowledge. People who just want to go to the rink and not dive into stats are going to argue there is no predictive value. Similar to baseball, there are going to be advantages to early adopters, and then every team will get a stats guy and the field will be leveled.

      Another area that’s impossible to predict: politics. And yet Nate Silver did predict the last US election, and the dinosaurs with their ‘gut’ predictions could not understand how they were so wrong.

      • shoobiedoobin - Apr 11, 2014 at 2:59 PM

        No. The word no is all I should need to say when you try to compare hockey to baseball.

        You look like you’re scared and defensive. Baseball can obviously be quantified more than any other sport, I mean c’mon just look at it. Hockey is the furthest thing from baseball.

  9. ndrick731 - Apr 11, 2014 at 1:45 PM

    These new analytics are just fantastic. They can tell you that the top four teams in the league have the best analytic rating and the worst four teams in the league have the worst analytic rating. It’s amazing.

    • sheckyrimshot - Apr 11, 2014 at 1:57 PM

      i just read about this incredible new metric called wins and losses. its suprisingly accurate.

  10. djshnooks - Apr 11, 2014 at 2:19 PM

    Thank you!!!

    THIS IS WHY BUFFALO WILL BE GOOD!

    Murray is huge on analytics and Pegula has dumped SERIOUS money into scouting and the analytics side of the operation…this will not be Edmonton all over again, they have the pieces and the people, as well as the money to get it right.

  11. muckleflugga - Apr 11, 2014 at 2:54 PM

    have to agree with shoobiedoobin

    reliance on statistics creates a common ground for folly … the stock markets can be brought to their knees by over-reliance on statistics by players using the same software employing statistics to automatically trigger buying and selling, producing communal catastrophe

    you see it at the track as well … twenty to thirty percent of winners are favourites, made so by people using the same data and analytical methods, roughly congruent with beyer speed figures for every horse … thank you for saving me the work

    with other favourites to discard, the field is narrowed and knowledgeable punters can concentrate on spot plays among remaining horses where odds are better and reward for risk produces better value … numbers cannot beat being at the track first hand to observe horses in the paddock and when they warm-up

    hockey’s the same … there are too many variables brought together in handicapping systems shared by anyone willing to buy the numbers, producing communal catastrophe

    analytics are of great value in producing general understandings for managers and coaches no doubt … without their use teams are playing at a disadvantage against teams better prepared … still, numbers do not measure heart and desire as it arises and grows in the moment

    and statistics cannot predict outcomes … who would work for a living if they did

    if you play the sports pools and bet hockey, don’t even consider making a bet until the conclusion of game two in each round … teams losing at home in either or both the first two games tend to continue losing … the inverse is true for teams winning on the road, not just by score but by superior and confident play

    reading body language of players on the benches and watching players interacting, even when trailing, is a big tell … teams who look confident and assured behave that way for a reason … they’re full of it and ready for a big run

    with a lights-out goalie and timely bounces and barrels of luck, teams are in with a chance in what is truly a crapshoot

    • salmon90 - Apr 11, 2014 at 3:19 PM

      and by “ready for a big run” you mean will probably out possess the opposition…y’know what a lot of these stats reveal.

    • grocerystick101 - Apr 11, 2014 at 4:22 PM

      And those teams and players that win in the playoffs and possess those intangibles will likely have their success reflected in these advanced statistics, as the numbers have shown.

    • hsnepsts - Apr 11, 2014 at 7:07 PM

      My colleagues and I invest 8 billion dollars in global stock markets, using statistics. We’ve been killing (usually) for over a decade. How? Because we know what we’re doing.

      You want to know a good stat from a bad one? Ask a statistician.

      You dont really get it. Your comment that “ther are too many variables in hoickey” illustates that. Good stats can tell you which stats are useless. That’s the point – turning data into real information – including where NOT to look.

      And about predictions – you dont need to be right all the time. In our business, if we are right 51% of the time, we make money. Its the same with a statistical analysis of sports. If you can help the GM make slightly better decisions over time, it adds up to a big win.

  12. muckleflugga - Apr 11, 2014 at 3:59 PM

    roger neilson was the godfather of video review and game analytics based on study of individual performance in real-time game situations

    his teams never won a stanley cup, but there is a statue of mr. nielson in vancouver

    waving a white flag in surrender

  13. muckleflugga - Apr 11, 2014 at 4:26 PM

    salmon90

    don’t tell me what i mean in effort to twist spin argument

    a big run begins in the moment … teams like the kings have won cups against odds and statistics predicting the contrary

    scotty bowman won a cup or two … he happily told anyone who listened that statistics are for losers … too right

    one might assume president’s trophy winners bring favourable stats to game one of the playoffs … how many of them have gone on to win cups

    the only statistic that matters is statistics evolving in the now

    • grocerystick101 - Apr 11, 2014 at 4:36 PM

      Actually, quite the contrary when it comes to the Kings, they are an example of advanced statistics at their finest. While standings-wise they were eighth in the Western Conference the year they won the cup, their Fenwick rating was fourth in the entire NHL during the regular season, an indicator that they were poised for a surprising postseason run.

      Again, these statistics are only one piece of the equation, but there is some clear correlation there. I don’t really understand why there is so much resistance.

  14. muckleflugga - Apr 11, 2014 at 5:17 PM

    grocerystick101

    true … but why only fourth … one would think the top ranking fenwick rating would have won the cup

    did the fenwick rating predict teams like dallas would gas their last game permitting kings to slink-in the back door … show me the mojo magic number cruncher who poured over spreadsheets to foresee that outcome

    the best thing proponents of statistics do is wave statistics … that’s what statistics are for … attempt to apply logic to the illogical

Top 10 NHL Player Searches
  1. C. Price (1755)
  2. J. Harding (1714)
  3. M. Staal (1580)
  4. J. Giguere (1517)
  5. J. Thornton (1463)
  1. A. Semin (1270)
  2. J. Spezza (1269)
  3. C. MacArthur (1246)
  4. A. Ekblad (1241)
  5. B. Bishop (1120)