BRIDGING THE GAP – My Take on Hockey Analytics Infiltrating the Mainstream

By Dan Kramer, @DanKramerHabs

I don’t claim to be an advanced hockey statistics expert. As an analytical type by nature, I’ve read up on Corsi, Fenwick, QoC, WOWY, and everything in between, primarily as tools to assess players I don’t watch on a regular basis, and/or to confirm what I’ve seen from what opponents of this school of thought term “actually watching the game.” But am I comfortable enough with the computations to whip out supporting arguments during dinner conversation? Just about as comfortable as I am when someone mentions the Israel-Palestine conflict around the water cooler at work. [read: not very]

As a product manager / marketer by trade, I know the value and importance of conducting research and crunching the numbers. When possible, you don’t want to make decisions on a whim, or decide between alternatives without a compelling case for either course of action. I’ve also seen how even proposals supported strongly by statistical analysis can crash and burn, often because someone failed to perform a simple reality check on the situation and apply sufficient common sense. Other times plans are altered in the face of numerical backing due to some outside factors that are tougher to measure with an Excel sheet and calculator.

What does this have to do with anything? Well the debate that has plagued Twitter, hockey forums, and blogs throughout much of the year – on the merits of so-called #FancyStats – has reignited of late, following the hiring of stat expert Tyler Dellow by the Edmonton Oilers, which in turn came after Sunny Mehta signed on with the New Jersey Devils, and Kyle Dubas became an Assistant General Manager with the Toronto Maple Leafs. And while these groundbreaking hires have been celebrated as vindicating victories by the advanced stats community and grown awareness of its existence, the truth is they are unlikely to greatly popularize their acceptance and use unless both sides of the discussion come to the table with different mindsets.

Reading through the many Tweets and posts, I couldn’t shake the feeling that I’d seen a situation just like this before. Finally, sitting in the office on a lunch break this week, it hit me, as I thought back to a teaching moment / development opportunity once shared to me by a workplace mentor.

“You, as an up-and-coming, fast-rising marketer, need to develop better ties with the salesforce. You know your stuff, and you’re eager to show your work to prove you’ve put in the time and effort to justify your conclusions,” she began. “But for them to accept you, you need to learn to communicate those same ideas in their terms. They’re afraid of someone they may perceive as smarter than them. They don’t want to look stupid, and they don’t need all your charts, graphs, and detailed explanations; that’s your role, and not theirs. They need you to uncondescendingly share the nuggets they can grasp and apply. Find your common ground.”

“Sure,” I said at the time. “That makes sense. I mean, I may live in Excel and Powerpoint today, but I used to work in sales (at another firm), so I’m sure I can use that to create a bond.” Little did I know that mindset was part of the problem.

I’ve always thought it funny – in the various organizations I’ve worked for – how marketing and sales don’t necessarily “get along.” They use different acronyms, have different metrics, and generally speak at each other rather than in a collaborative way. When some initiative that should have been considered a joint undertaking doesn’t go according to plan, the sides are quick to point fingers at one another. All of this despite the fact that, theoretically, both sides are out to achieve the same thing – to maximize the profits of their corporation, or succeed at some other predetermined objective.

Consider advanced stats. Both the calculator-touting new schoolers and the eye-test-only traditionalists seek to assess players. Both want to identify under-the-radar free agent targets, proclaim winners of trades, and/or champion the next wave of promising prospects. One might think, then, that the two would be aligned more often than not. Yet this isn’t necessarily the case.

I was close in my line of thinking. It is about finding common ground. It’s just not about using that common ground to prove that your way is “right” or of superior quality. The mass media, not unlike your typical salesforce, are largely uninterested in the charts and graphs that fill lengthy, detailed articles produced by the fancy stat-ers. To most of the public, the numbers just aren’t sexy; instead, they bring back less-than-fond memories of high school math class. The advanced stat community, not unlike the marketers, are then left talking mostly amongst themselves, in the holier-than-thou space of needing to extol every finding that led them to some finish line. The land of one-up’ing. Of being “smarter” than the rest.

“Oh no, not I,” may say the seemingly less pretentious of the lot. “I watched hockey before these stats were a ‘thing.’ I used to be just like the other side, but now I’ve supplemented my game-watching with a new piece of information. I’ll teach them how to do the same.”

Did you catch it when I spelled it out that way? This is the unintentional crime I was committing in my “oh I used to be one of them” pretense. That the other side could – or even SHOULD – make the same leap I had. It isn’t so much that they aren’t capable of it, but rather the underlying assumption that having an intimate understanding of advanced stat numbers represents a positive evolution as an observer of the game; evidence that you are of a superior intellect, or that you’ve progressed to a higher state as a player evaluator. Looking at things in this way helps to understand why no amount of “dumbed down,” simplified, or thoroughly explained graphs will be sufficient in converting the uninterested.

Continuing to attempt to reach the masses by those means would further imply that so-called traditionalists don’t or can’t evolve in their understanding of the game. This erroneous thinking only grows the chasm that separates the two schools. They aren’t “dumb sales guys” of the 1960s. They are in the field daily, watching, interacting, engaging, and re-evaluating their previous thoughts. They are building on a wealth of accumulated experience, while layering on all the latest trends. Hockey analysts HAVE progressed in their thinking, just as they don’t judge goaltenders by the same standards as we did in the 1970s. Every new player, new coaching system, new tactic evolves the way they evaluate an individual’s skill set. Does some archaic old boy’s club groupthink that rejects change in any form exist in some circles? Certainly, but this is likely equally as small a minority as those who believe in strictly looking at the numbers and ignoring anything else.

The truth is, the more hockey one watches and/or researches, the more their understanding of the game and its players grows. It’s not unlike business/economic writer Malcolm Gladwell’s Outliers theory by which it takes 10,000 hours at something to become a true expert or professional. One will learn the differences between what makes a good player and what is more likely to be a poor acquisition by comparing experiences and learnings accumulated over time. If anything, those fancy charts and graphs may save one time by allowing the processing / simplification of large amounts of real-time data all at once, but certainly they are not synonymous with a better grasp of player evaluation.

How do we bridge the gap, then, and put an end to the nasty Twitter exchanges and vengeful, sarcasm-laden articles? Rather than proving how sophisticated and reliably predictive the numbers are (though of course there is still place for all of that in supporting documentation), the key is in the outcomes that both sides can understand. Both sides can agree that a player’s ability to generate attempts on net speaks to his skill. Someone who favours either evaluation toolset judging a player’s ability to maintain puck possession isn’t anything novel. Once these ideas become as household as “skating ability,” “shot power,” and “stickhandling” that you can find quantified in every EA Sports game on any platform dating back to the 90s, then we can gradually layer on obscure terms and statistical measures to visually compare players.

Simply put, going from goals and assists to Corsi and Fenwick is straying too far from what is comfortable for the casual fan or even run-of-the-mill analyst. It’s not that they’re stupid, or couldn’t understand it if they wanted to; it’s just not part of what they enjoy about the game of hockey, and reading detailed statistics won’t necessarily increase their understanding of a player they’ve watched repeatedly. Why, notably this summer, have teams increasingly begun to turn to those with a good grasp of these statistics? For one, the concepts are likely sufficiently socialized within the organizations making these hires to be taken for what they are, essentially summarized observations. The mass public outside those bubbles isn’t there yet. And for another, it needs to be noted that while the hires are getting significant press, the keys of the castle aren’t exactly being turned over to a new breed, and remain squarely in the pocket of the men who were – and still are – actually in charge. The movie Moneyball popularized the idea of a behind-the-scenes underling truly calling the shots, but such a scenario is highly unlikely to play itself out in the National Hockey League. The new guys will be just that; another voice in the room with perhaps a different perspective to add to – and not supersede or replace – those that were already there.

What am I saying here? In essence, without aiming to demean the amazing work that those in the community have done, it needs to be made clear they are still watching the same game – and looking for many of the same things – that those in the know have followed for years. The path to mass understanding and acceptance of this wave of statistics is getting that it represents a change or formalization rather than a true evolution. It isn’t the discovery of a glitch to score trick goals like one might exploit in the aforementioned EA Sports games, but is a numerical expression of what many scouts and reporters have spoken to in assessing players for generations. Again, I aim not to diminish the work, but to show it for what it truly is: the invention of a way to track what our eyes already tell us, enabling an analysis over greater sample sizes or a comparison of different players and situations.

Times are changing when it comes to the acceptance of these new metrics, no doubt, but these hires are likely more a symptom of the change rather than the driving force behind it. What will drive it is continued exposure to some of the concepts in easily relatable ways – such as TSN (or Sportsnet) broadcasts including shot attempts next to shots on goal – which don’t even require pretty, hoity-toity graphical illustrations. It isn’t because the analysts and fans lack the intelligence to take to these concepts immediately, but they simply need time to associate these foreign words to the spot they already know and love. So in the meanwhile, leave the graphs at home. We’ll believe your findings without them.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s