From Puck Luck to Big Data: The Advanced Analytics Revolution in the NHL (Undergraduate Journalism) / by Olivia Reiner

Ron Hextall (far left), the general manager of the Philadelphia Flyers, is among several NHL bigwigs that hired advanced statistics analysts over the summer to work for their organizations. Other teams that hopped on board include the Toronto Maple …

Ron Hextall (far left), the general manager of the Philadelphia Flyers, is among several NHL bigwigs that hired advanced statistics analysts over the summer to work for their organizations. Other teams that hopped on board include the Toronto Maple Leafs, the New Jersey Devils, the Edmonton Oilers and the Carolina Hurricanes.

When it came to creating a user-friendly website and vibrant graphs detailing win expectancy in any given game, nobody did it better than Darryl Metcalf, founder of Extra Skater, a popular watering hole for the hockey fan community.

Metcalf had such a highly regarded hockey analytics blog that he was forced to shut down operations for good. Metcalf wasn’t sent to the penalty box for unbloggerlike conduct; he was scooped up by the Toronto Maple Leafs to work his analytic magic for the wealthiest franchise in the NHL.

“Extra Skater was a godsend for those of us that liked this stuff,” said Mark Lazerus, the Chicago Blackhawks beat writer for the Chicago Sun-Times. “Extra Skater was so easy to use that [writers] kind of depended on it.”

For Lazerus, advanced statistics enhance the stories he writes for the Sun-Times. While he avoids using technical units of measurement like “Corsi” and “Fenwick,” he assesses their meanings in simpler terms to illustrate a dynamic game in stagnant print. Corsi, which measures the sum of shots on goal, missed shots and blocked shots, and Fenwick, the same quantifications as Corsi minus blocked shots, helps Lazerus explain the importance of puck possession in a given game. Since Extra Skater’s termination, he has to look elsewhere for his data.

The Big Data revolution in professional ice hockey started as a grassroots movement among a niche group of curious fans with a knack for calculations. Each particular measurement evaluates a number of factors to help predict future success of individual players and teams or assesses what affects a win. Over the summer, several NHL organizations saw the value of this data and created official advanced statistician positions, often filling those roles with former bloggers.

“Four or five years ago, I stumbled onto an online community that was talking about hockey analytics and got myself interested,” said Eric Tulsky, who was a blogger prior to becoming an analytics consultant for the Carolina Hurricanes. “It’s a neat thing to go from just saying what I think teams ought to do to actually helping a team decide what they’re going to do.”

Although Tulsky’s articles on SB Nation’s Outnumbered are still online, the work and research of Metcalf and other former bloggers like Tyler Dellow, who now works for the Edmonton Oilers, have disappeared.

“The problem is a lot of this is proprietary information in proprietary systems,” said Paul Kennedy, a former media relations spokesman of the Chicago Blackhawks. “The really good statistics are not typically going to find their way into fans’ hands because teams believe that they give them a better chance to win.”

Taking away these brains from the public forum is troublesome for the community, according to Stephen Pettigrew, founder of Rink Stats and contributor to the Wall Street Journal, DeadSpin and FiveThirtyEight. Professional organizations stifle creativity when advanced analytics proponents lose an intelligent public voice.

“Most of the cutting edge stuff is going on at the team level,” said Pettigrew. “The problem is that because it’s being done by teams, we don’t really know what the cutting edge is. Teams don’t write research papers about what they’re doing. If they find something that works really well, they’re not going to write a blog post about it.”

However, websites like War on Ice and Hockey Analysis have popped up in the absence of shut down blogs to fill the public research void. In late February, the NHL launched an “enhanced statistics” page for public use in addition to the proliferation of new sources.

“If it comes from the NHL itself, you can assume that it’s correct because it’s their league, it’s their information,” said Lazerus. “I think it legitimizes the stats when those guys do get hired by teams because clearly they’re doing something right.”

Despite the vast quantity of blogs, Kennedy stresses the importance of quality of information. Dellow, he said, worked on innovative “thought experiments” by ruminating on what he observed in the game, using the online community to develop his ideas. That, according to Kennedy, is what the public analytics world currently lacks.

“Who is going to be the next person publicly to do some of that brainstorming out in the open that fans can participate in, that fans can be a part of, that they can see the process of where it comes from?” asked Kennedy. “There is always going to be someone else, I hope. I hope there will always be someone else.”