Analytics still a hard sell in the NHL

"I can't wrap my head around why the analytics guys hate the old-school guys and the old-school guys hate the analytics," says former Oilers coach Dallas Eakins, an analytics proponent. Jonathan Kozub/NHLI/Getty Images

When the Edmonton Oilers hired Dallas Eakins in June 2013 to kick-start their youth movement, he was seen as the archetype for the new breed of NHL coach. Eakins, 46, was a former NHL player, a successful American Hockey League head coach and well-versed in the world of advanced stats.

In fact, Eakins considers himself an analytics proponent. Last summer, Eakins championed the hiring of prominent analytics blogger and vocal Oilers critic Tyler Dellow for a role in hockey operations. With Dellow's input, Eakins tweaked the way the Oilers played in order to improve their Corsi, an advanced stat that measures possession. Edmonton saw a significant improvement in Corsi percentage under Eakins from last season to this season.

The wins column was another story. After the Oilers lost 15 of 16 games, Eakins was fired in December, just 18 months after he was hired. In one of his first interviews since leaving Edmonton, Eakins spoke with ESPN The Magazine's Craig Custance about the challenge of selling analytics in a sport that remains resistant to change.

Custance: What's the hardest part about selling analytics in hockey?

Eakins: I don't even know if "sell" is the right word. The really interesting thing about advanced stats and analytics is that they're just another column of stats. It's the same as saying, "This guy has this many goals," or "Your power play is running at this percentage." When I was in the American Hockey League, I was taking the time to read about [advanced stats]: What is Corsi? What is Fenwick? What is PDO?

Custance: How accepting did you find hockey of using those stats?

Eakins: When you talk to people in hockey, and even fans, there's such a divide. You're either for these analytics or you're against them. For the life of me, I can't wrap my head around why the analytics guys hate the old-school guys and the old-school guys hate the analytics. Why wouldn't you look at [advanced stats]? I don't understand why you would not at least look at them.

Custance: The advanced stats suggested there was progress being made in Edmonton. Is that what you saw internally?

Eakins: I got fired from a job and I thought we were really moving -- if you wanted to look at it stats-wise -- in the right direction. But in the end, the only stat that counts is if you win or you lose.

Custance: How did you convince the Oilers to hire Tyler Dellow in their analytics department after he'd been so critical of the organization as an analytics blogger?

Eakins: I was unaware of the background with Tyler. I met him at a coaching clinic where he was a presenter. After the clinic, the coaching staffs went out to dinner, and I sat next to Tyler. He seemed like an intelligent kid, really passionate about the Oilers. I thought that within five years every team would have an analytics coach or person on their staff, for sure. I said: "What are we waiting for? Are we going to wait until year four and see who is left?" That's when I learned he'd been highly critical of the Oilers. But how is he supposed to be positive when, over the last five years, there have been four coaching changes and the team finished 23rd or 24th in the league? They were one of the bottom teams in goals for, shots for. What is there to be positive about?

Custance: What are the challenges in making sure an analytics hire has a voice that's being heard by others in the organization?

Eakins: It was simple with my staff. I was the head coach. When the head coach says something, they're going to listen. I let my whole staff know we hired this guy, and we hired him for a reason. He would have a voice at the table, no different than any other person on staff.

You might sit with five or six coaches and try to determine which way you're going to forecheck. You might have four different opinions. Tyler would bring the analytical side to it. Sometimes we would go with it, sometimes we wouldn't go with it. He was just another voice at the table.

Custance: The Oilers' CorsiFor Percentage improved from 44.3 percent last season to over 50 percent this season before you were fired. Do you credit that to system changes, personnel -- or both?

Eakins: If it made sense both ways to make a change -- with our eyes, our experience and the numbers -- then we would switch it. We switched a few things -- very, very subtle changes, but they were paying off. I'm not sure how much better we were in terms of players than we were last year. It could have been apples to apples. But we were able to change some things.

Custance: When you're implementing systematic changes, is part of selling it to the players explaining the analytics, or does that just get in the way?

Eakins: This generation of players reads everything. When I was a young player, if I wanted to read something I had to go buy the newspaper and pay for it. Now you go to the computer. You Google your name. You're reading everything. The one thing we thought we needed to do was make sure the players understood [analytics]. They all know goals and assists, points and power-play percentage. We wanted to make sure that they understood what Corsi and PDO were. We wanted to be very open-book about it. It's how you have to be in life and as a coach. You have to be up front and honest. We took the approach of: "Hey, this is how many points we need to make the playoffs. This is how many points we think our power play should be at. This is what we think our Corsi should be." That's when it was great to have Tyler on staff. He was able to explain and answer questions more in-depth than I could. He could nail it right away.