When the Arizona Coyotes hired 26-year-old John Chayka as their new general manager, it was headline news across the NHL. It was also the latest indication of hockey’s evolution into a 21st century sport with the influence of analytics.
Being the youngest General Manager in pro sports history, besting the Toronto Maple Leaf former GM Gord Stellick by over four years, is what made the headlines around the hockey world. The Hockey Writers own Neal McHale wrote about how his hiring has opened the door to the old-boys network of NHL management. You can read his great piece here – ‘Chayka Hiring Opening the Secluded Door‘. What caught my attention was the path Chayka took on his way into the top spot of a NHL front office – the new path of advanced analytics.
Before he was hired by Arizona last season as Assistant General Manager/Analytics, he was Director of Hockey Operations for Stathletes Inc., a company that he co-founded, that touts itself as “Professional hockey’s deepest performance data & analytics”. The Coyotes official announcement described the company this way:
Stathletes is a hockey analytics firm that tracks data through an intensive video analysis process and breaks down the game to provide objective insight into player and team performance tendencies.
Many teams have strongly embraced the analytics movement in hockey, hiring former statisticians and mathematicians for key management level positions. News even broke on Sunday of front office changes with the Florida Panthers that involves a promotion for their own number cruncher, Eric Joyce, to Assistant GM. Chayka, though, represents the first ‘Corsi Guy’ to reach the corner office. His hiring proves that the use of advanced statistics and data driven decisions are now firmly embedded in the NHL culture, and it’s upper-level management.
Let me be clear right from the start. When advanced analytics started making its way into the lexicon of hockey news and opiniosn I wasn’t a fan. I was one of the old guys railing against these young pups, with their numbers and charts, who thought they could break down a player’s performance in any given situation with a bunch of numbers. My knee-jerk reaction was to dismiss the data as mumbo-jumbo, and the analysts as the nerds I knew in high school who spent their time in front of the Commodore 64 instead of out on the ice playing.
I was REALLY wrong.
I grew up in a world of hockey long before the Corsi and Fenwick scores of today. The most advanced stats we ever had were goals, assists and penalty minutes, and goalies that let in fewer goals than the other guy. The most in-depth analysis of a player was whether he hustled on the ice, handled the puck well, could give and take a solid body-check and was a team player. Back then the coaches and general managers had all the basic numbers to evaluate a player’s performance, but after that it was much more personal. It was what they saw, heard, felt and sensed from the player and the team. It came from years of experience and knowledge, and many times came down to instinct or a ‘gut-feeling’.
In other words, for decades, I was this guy when it came to advanced stats in hockey:
Tomorrow, we wake up in the 1st day of post-Corsi era. Life will be good, we will all be beautiful, hockey articles will be poetry #PHIvsWAS
— Slava Malamud (@SlavaMalamud) April 23, 2016
Then, earlier this year, I started writing about hockey and the Carolina Hurricanes. Although, I still held the opinions of my Russian friend, my articles and editorials needed to include ALL the information about the teams and players I was writing about. I couldn’t have current and relevant pieces about the ‘Canes without using the information and stats that the readers wanted. So I sucked it up, got on Twitter, and started searching for the best sources of this new-fangled stuff. That’s when I started running into this:
I couldn’t figure all it out. I had to have three windows open on my laptop so I could read a stat, flip over to the page that described the stat I was reading, flip back to my article to try and write about the stat I just read, flip back to the stat description to make sure I used it correctly, flip back…you get the idea. The ‘old-guy brain’ that hadn’t dealt with advanced math since my senior year was on tilt. I was ready to throw up my hands and join the fight against the ‘Corsi Mafia’ like my friend Slava. Then I had a minor epiphany.
One day I was doing some of the usual reading and research on the Hurricanes’ website when I came across a post-practice video of Head Coach Bill Peters. He was talking about the young players he was having to evaluate as the revolving door between the Hurricanes and the Charlotte Checkers kept his roster changing all season. He was asked about various players and how their development was progressing, and if they were showing signs of NHL caliber talent.
Now, to be completely upfront, I LOVE Bill Peters. I love him as a coach, a leader, a teacher and as the “straight old-fashioned Alberta Redneck” that he is. Peters is old-school hockey coaching. He wants a fast playing, feet always moving, 200-foot focused team. He doesn’t want his guys walking in to walls…he wants them running in to them. If they make a mistake on a shift, but are working hard when they make it, they aren’t benched. They’re sent back out to try it again and do better. Practices should be as fast and hard as games, or you might as well be at the public skate. It’s a simple ‘play hard or you don’t play’ coaching mentality learned from the greats like Scotty Bowman, perfected as an assistant to Mike Babcock, and familiar to anyone who grew up playing hockey in small Northern towns.
Then I came across the Hurricanes new Hockey Analyst Eric Tulsky, and started reading about what he was bringing to the team. The Philadelphia born, Harvard and UC-Berkeley educated, Ph.D. in chemistry, Naval Research Lab nanotechnology expert had been been publishing his advanced hockey statistical analysis since 2011. He had contributed to NHLNumbers.com, launched his own website, Outnumbered, and his hockey analysis had been published in The Washington Post and fivethirtyeight.com.
I wondered how a number-crunching whiz kid with a background in DNA sequencing and solar panel technology could help two old-time hockey minds like General Manager Ron Francis and Bill Peters. As the Hurricanes’ season progressed, I slowly learned how his expertise in breaking down the advanced stats and data was helping Peters grade a bunch of rookie players that he had no previous experience with. During post-game and practice interviews Peters would reference Tulsky’s work when he was asked about a player’s development or performance. He never went into the details of the data or numbers, but he would use the statistical evidence that Tulsky provided as proof of what he saw in the player.
The light went off for me. I didn’t have to go back to school and learn advanced calculus in order to follow or write about the ‘new hockey’. I didn’t need to become an expert in all the levels of advanced analytics to be able to write about, and form an opinion, on a player. Analytics in hockey aren’t intended to replace what has worked for generations of coaches and general managers, but to enhance and add to what they already know about a player or a team. They are meant to reinforce what the coach is seeing and feeling, or help explain an issue that the coach just can’t figure out with his eyes and ears.
These quotes from Francis and Peters, during the Hurricanes end-of-season presser, encapsulate what I’m trying to say. When asked what Tulsky brings to the table and how that affects decision-making, Francis and Peters said:
Francis: “I think he is an extremely bright mind, and he thinks a little bit outside the box. My interpretation of analytics is that it’s a balance and check. What I may be seeing or what I think I’m seeing with my eyes, it gives me a chance to back up what I’m seeing or question what I’m seeing. He does a lot of neat things.”
Peters: “I think it’s been an unbelievably valuable asset to our organization. I enjoy the conversations with Eric and the information he puts out to us as coaches. It’s very thought-provoking and helpful. He makes you better. It makes you think and think through different scenarios and situations. It’s all about finding ways to be more successful. I embrace it.”
In our all-or-nothing opinionated world, where you either agree with me or you’re just wrong, the use of advanced statistics and analytics in the hockey world has created many heated arguments. At the extremes the old-school guys dismiss the numbers as the mumbo-jumbo and the Corsi-lovers ignore gut-feelings and instincts.
I have learned that the best situation is a nice balance of both. Not everything on the ice can be quantified into columns and graphs, and on the other side actions and results need to be qualified in order to be understood. When you’re evaluating a player and everything is equal, you need another source of information to help make the decision. When the size and effort you see in two players is the same, you can turn to the advanced analytics. When the statistics and charts for two players are the same, you can fall back on your experience and instincts.
The smart hockey teams will find the balance between their instinctual coaching and management staff, and their data-driven number-crunching geniuses. The Arizona Coyotes are simply the first team to put one of the geniuses in a high profile decision-making position. I will be watching closely, and just between you and me, hoping Chayka and the Coyotes are successful. Don’t tell my Russian friend Slava.
Grew up in the ‘hockey town’ that launched the careers of Bob Gainey, Roger Neilson, Scotty Bowman, Chris Pronger and Steve Yzerman, to name only a few – Peterborough, Ontario. Spent a misguided life in the world of politics. (Sorry Mom) Trying to atone for those sins by publishing the best hockey stuff available. You can email me directly at Dean@TheHockeyWriters.com. I don’t accept emails written in Sanskrit or asking me to help you transfer your family funds from Nairobi…even if you are a Prince. Thank you.