Analyzing the smart game
The raging debate on hockey statistics
Benjamin Disraeli, the great British parliamentarian of the 19th century, reputedly divided lies into three categories: “lies, damned lies, and statistics.” This might have been the original salvo fired into a battle that continues to rage in our own century. It’s fitting then, that Athabasca University’s (AU) MBA in Hockey Management program brings a learned perspective to the ongoing debate about statistics as they pertain to the business of sport. Dr. Rodney Paul, professor for the program’s Hockey Operations course, offers a “numbers week” that delves into advanced analytics and how they can help NHL organizations evaluate player and team performance, among other functions. He leads his students through an investigation into the data currently available to hockey management, where the data can be useful, and where it can be misleading.
Insiders vs. outsiders, nerds vs. old pros
This reasoned, academic approach suggests that the debate about statistics doesn’t have to be an either/or proposition, unlike Hollywood’s depiction in the Oscar-nominated, 2011 hit Moneyball. Against the objections of flinty-eyed, veteran baseball scouts, a nerdish, Yale-trained economist used “sabermetrics” to give the penurious Oakland Athletics a player-evaluating edge against their wealthier competition. (Michael Lewis’s 2003 book of the same name told the truer, more nuanced story.) Since the early 2000s, advanced analytics have gained wider acceptance in the Major Leagues, but their incursion into the hockey world has been slower. Dr. Rodney Paul acknowledges that even in 2017 there is still resistance in some professional hockey organizations regarding the newer statistics that are now used by many analysts, “There is some natural friction between old school methods and new data-based evaluation. In a perfect world, you find those synergies [involving both models].”
This natural friction still provides fodder for bloviating in the hockey blogosphere. For example, in a July 2017 post barking “Shots fired!”, Cult of Hockey blogger David Staples set a former NHLer’s comments about the role of toughness in the game (by implication unquantifiable) against the numbers-based analysis done by “hockey nerds” (Staples’ self-inclusive term). After much back and forth considering the merits of both sides – new stats or intangibles like team cohesiveness – he left the issue framed in a bi-polar paradigm.
“There is some natural friction between old school methods and new data-based evaluation. In a perfect world, you find those synergies [involving both models].”
– Dr. Rodney Paul, Athabasca University’s MBA in Hockey Management program
But in the real, results-driven world of the hockey business, that paradigm of numbers nerds versus savvy old pros, while not gone completely, is morphing into something different. Now, the synergies of which Dr. Rodney Paul speaks are providing value for many NHL franchises. Brian Burke, co-founder of the Business of Hockey Institute, (AU is the educational provider to the Business of Hockey Institute) is also President of Hockey Operations for the Calgary Flames. As one of the NHL’s arch-insiders and old-timers, Burke is nonetheless a hard-nosed proponent of evidence-based decision making and proud of the Flames’ investment in the new science of analytics: “We have the best analytics guy in the league [Director Chris Snow].” Burke explains that Snow is an important voice at the table in the Flames’ operation; working with scouts, coaches and management on player evaluation, with management on contract preparations and negotiations, and with the coaching staff on tactical preparations from game to game.
Other NHL organizations have gone deeper with the numbers game. In 2016, the struggling Arizona Coyotes hired then-26-year old, John Chayka, a statistics wunderkind, to be their general manager, raising eyebrows throughout the league. While the jury is still out on that precipitous move, the Toronto Star dubbed 2016 “the summer of analytics,” reporting that 20 of the NHL’s then-30 teams (there are now 31) listed at least one analytics-related employee in their staff directories.
In 2017, it’s hard to tell from scanning NHL websites which organizations have invested more in analytics than others. Some show multiple employees dedicated to the function, others none. It’s worth noting that Brian Burke’s fierce rivals and immediate neighbours to the north, the Edmonton Oilers, identify no analytics specialist on their website, yet they have several staff members devoted to comprehensive statistical analysis.
Smartest game, hardest to analyze
The question that continues to dog the NHL, and hockey in general, isn’t so much about whether advanced stats can help measure player and team performance, but which stats can and should be applied. American journalist Adam Gopnik indirectly addresses this strategic challenge in his thought-provoking essay “Why hockey is the smartest game in the world” (commissioned by the CBC for the 2011 Massey Lecture series).
Employing mathematical game theory, and psychological concepts such as spatial intelligence and situational awareness, he compares the frozen game to other major team sports, “Hockey approaches a more perfect balance between planning and reading, idea and improvisation, than any other sport.” He breaks down some of the great goals of hockey lore, suggesting that each one “is the result of a plan and history unknown to or beyond the control of the opposition, shared among the players through their common spatial intelligence, each taking place at such high speed that the plan is invisible to all but the tutored eye.”
One can infer from Gopnik’s almost lyrical praise of the game that it can be subject to empirical analysis (“planning and reading”), and yet because of its speed and team chemistry (“common spatial intelligence”), it can at times defy attempts to quantify its working parts.
Brian Burke weighs in on this conundrum simply and effectively, “Baseball is uniquely suited to statistical analysis because it’s a series of identical, repetitive events. Only the pitches change. Hockey is much more random.” He says the physical element of the game is harder to quantify (even with stats on hits given and taken) and throws in his favourite term ‘truculence’ – “How do you measure that?” That’s why Burke’s Flames use analytics as one tool for evaluative purposes. Two others, which Burke ranks higher in his preference, are the “eyetest” (watching a player perform on the ice over a period of time), and character checks (references and player interviews).
The AU perspective
Dr. Rodney Paul believes the stats he covers in his “numbers week” can support Brian Burke’s higher ranked tools. He hopes the day will come when the experts devise statistical methods to assess even the random, capricious elements of the game. For now, though, he and his students, and the pros who work within the NHL and its constituencies, will rely on current arcana such as Corsi numbers, Fenwick numbers, the Royal Road, green goals, red goals, WOWY (“with or without you”), zone start percentages, and on and on. Today’s hockey fan maybe conversant with some of these terms, but hockey operations pros must be able to research them (through many independent databases), in some cases generate them, interpret them, and provide sound strategic counsel to senior management based on their expertise with these data.
At least two of AU’s MBA in Hockey Management students are well versed in advanced analytics. Shane Malloy works daily with analytics, generating detailed data reports on every NHL, AHL, and CHL player for Electronic Arts, the company that produces wildly popular pro hockey video games. The players featured in these games make their virtual moves based on the data provided by Malloy. He’s also authored an homage to the work of amateur hockey scouts, The Art of Scouting. His take on analytics? “[It’s] been around in hockey for 10 years or more, but it’s still in its infancy. We’re going through the process of learning what metrics are most useful.”
“The 24/7 sports news stream has pressured the business into finding ever more things to talk about. Numbers and their proponents feed some of that need.”
– Kerry McGowan, student in AU’s MBA in Hockey Management program
Kerry McGowan is a fellow cohort member of Malloy’s, and in the same vein, he’s making use of analytics in an ancillary part of the hockey business.
He’s a shareholder in the Nation Network, a series of NHL fan websites that make extensive use of bloggers who draw from analytics databases to provide commentary for hockey enthusiasts. McGowan is also the owner of an oilfield services company, and like Burke, he’s a believer in stats and evidence-based decision making. “It helps you avoid breathing your own exhaust.” As for stats and the Nation Network, “The 24/7 sports news stream has pressured the business into finding ever more things to talk about. Numbers and their proponents feed some of that need.”
A synergistic future?
Kerry McGowan and Shane Malloy know both the marketing value and the strategic value of advanced analytics. The new numbers are obviously a vital and growing part of many facets of the hockey business. The combination of forward-thinking hockey executives, and academics like Dr. Paul and is students, will drive this growth into the future. The professor speculates about a time when the synergies he seeks may possibly help measure intangibles like truculence and team chemistry. He says that with a large enough database more patterns will emerge, and more collaborative use of different tools in hockey operations will continue to improve decision-making. And at least a few graduates of AU’s MBA in Hockey Management will play a role in making the world’s smartest game even smarter.
Originally appeared in the Fall edition of Athabasca University’s Faculty of Business Connected magazine.
