Coyotes up their analytics game with stats guru Chayka

John Chayka doesn’t expect to find the holy grail of fancy stats in his new role as Coyotes assistant general manager for analytics.

John Chayka

"There is not one measure that will embody the value of a player," Chayka said. "There are different metrics and they have different strengths and weaknesses. It’s about finding out what the metric is telling you and then trying to answer some of the questions it creates."

Chayka isn’t wasting any time getting rolling. He spent last week at the NHL Scouting Combine evaluating and interviewing players with general manager Don Maloney, assistant GM Darcy Regier and members of the Coyotes scouting staff.

This week, he’ll make the trek to Arizona to start house hunting while helping the staff prepare for the NHL Draft on June 25-26 in Sunrise, Fla., and the start of free agency on July 1. 

Since hanging up his skates after playing parts of three seasons of junior hockey, Chayka has been focused on growing Stathletes, the company he co-founded with a friend to manually track games through exhaustive video analysis.

While few NHL players profess to know much about analytics, and while many profess skepticism, Chayka isn’t concerned about getting buy-in from the Coyotes when he presents his data.

"I actually started in the player development industry working with individual players to understand players’ strengths and weaknesses," he said. "I used a lot of video, and I saw that players actually preferred the data we gave them because it was quicker and easier to understand. I found I could use certain benchmarks to get a point across."

Chayka, 25, said he has had offers to join other NHL teams in the past.

"I never felt I was ready for a position like this," he said. "But Arizona is a place that I wanted to be because it’s an exciting opportunity. The biggest change agent is necessity, and it’s necessary in Arizona that we think differently and more creatively based on the economics of that marketplace.

"You talk to people around the league, and Don (Maloney) is so well respected. That opportunity, along with working with Darcy (Regier), who obviously has an interest in this data, and Dave Tippett, one of the best coaches in the NHL, made me think it was the right time for me on a professional level."

Much of the criticism of the current NHL analytics stems from the fact that they are collected from the league’s often-unreliable live statistics reports. There is notorious variance in the tracking of shots, shot attempts, hits, giveaways and other statistics from building to building. 

It’s possible to account for some of that variance with margins of errors and other measures, and the current crop of advanced statistics paints a useful picture of what’s happening on the ice. At the same time, Regier, Tippett and the Coyotes staff are looking for more reliable data, reasoning that bad data in means bad data out.

"It’s obvious that in most games, if you have more shot attempts, you probably possessed the puck more, so there is a correlation between Corsi and possession," Regier said recently. "But those stats don’t measure possession. They measure shot attempts. You’d like something a little more solid."

Tippett has been outspoken in his criticism of the two best known advanced statistics: Corsi and Fenwick, derisively referring to them as "Corski" and "Fenski." That has led some to believe that Tippett isn’t an advanced stats disciple.

"I think to the average fan, these stats can be dangerous because everyone’s looking for that one option, and I don’t think that’s ever the way to evaluate anything," Chayka said. "It’s how these numbers work together to maximize a team. Specific players bring specific attributes but might have weaknesses. It all affects the holistic view of how you build a team.

"There’s a mountain of data out there, so you have to be careful when using it and analyzing it and dissecting it because there’s a lot of noise. There are a lot of things we’ve tracked over the years that, intuitively you would think would be useful data but aren’t at all. I’m not trying to maximize one specific thing, so you can’t evaluate a player based on one thing. I’m trying to build balance and equilibrium to win games." 

Follow Craig Morgan on Twitter