When the request for an analytics breakdown on Taylor Hall came in, Sai Okabayashi didn’t think too much about it, mostly because he didn’t really think the New Jersey Devils had a shot at acquiring him.

Okabayashi was part of the New Jersey analytics team led by director Sunny Mehta, and those requests from GM Ray Shero would sometimes come via a text message, paired up with queries about two or three other players. Often, nothing came of them.

So when he saw a request for the usual analytics package on Hall, he wasn’t necessarily thinking a monster trade was coming.

“It’s like, ‘Okay, I’ll give you information about Hall, but why bother?’ The guy is 99th percentile of every possible category,” Okabayashi said when we chatted on Monday.

He estimates that the first time he was asked for the analytical breakdown for Hall, it came about a week or so before the trade for Adam Larsson happened. And yes, the analytics on Hall were great.

“Oh yeah, yeah,” Okabayashi said. “Outstanding.”

Okabayashi doesn’t believe anything he provided to Shero swayed the deal. He thinks it was happening anyway, because Hall exactly fits the kind of player that the Devils are acquiring right now in a rebuild without tanking. They try to find talented players that other teams are undervaluing for some reason. The Oilers were loaded with young forwards, and they were signing Milan Lucic. Someone had to go.

The Devils took advantage.

But just being a part of the process was a thrill for Okabayashi, a die-hard hockey fan. As a member of the Devils' front office, he would attend games in New Jersey and take a moment to soak it all in. He suspected this wasn’t going to be a lifelong journey with the team, so he savored it while it lasted.

His suspicions came true when his wife was offered a great job in Seattle, one his family couldn’t pass up. So, after making sure the Devils were in good shape to start the season, he left his job in hockey to join a startup out West.

“Great guy,” texted Devils GM Ray Shero. “Wish he stayed.”

Those of us on the outside learned about his departure the same way we learned about his hiring by an NHL team -- from the activity on his fantastic analytics site shiftchart.com. Specifically, it started working again.

The gap in his posts on Twitter -- the last coming on Aug. 26, 2014 followed by a rebirth on Nov. 4, 2016 -- reveals the timeline of his employment with the Devils.

That ShiftChart is back and operational for hockey fans to see is a great thing; it’s one of the few places where there’s a visual for the line matching that goes on during games by NHL coaches.

Okabayashi’s return to the normal world with the rest of us is also an opportunity to hear about what it’s like on the other side of the curtain, where how analytics are used is a well-kept secret.

When he was first hired by the Devils, an ownership-driven hire, his focus was on building an internal database infrastructure and tools for the Devils to use in order to create reports for the front office and coaching staff.

He also spanned two front offices -- the first headed by Lou Lamoriello and the second by Shero. Over those two regimes, he saw a very different style and use of the analytics team.

With Lamoriello, it was very much a one-way street.

“I think he was very intimidating and at the same time respectful,” Okabayashi said. “I was sad to see him go. When Ray came in, it opened doors to have more conversations with coaching and with scouts and allowed us to have more impact.”

That’s when his role started to change. Under Shero, the analytics team was welcome to have conversations with scouts, coaching staff and front office staff.

Often in hockey, it’s painted as a battle between the old-school people and the analytics teams, but in guys like Shero, assistant GM Tom Fitzgerald, scout Bob Hoffmeyer and coach John Hynes, he found people willing to share information and engage in discussion.

When analytics experts started to get hired a couple of years ago in hockey, there was a concern that they wouldn’t be integrated into the decision-making. There were people who thought they might be locked in a room with computers as token hires to satisfy ownership or the fans.

That wasn’t the experience for Okabayashi.

To make it work, it takes an open-minded front office and it also takes an analytics person who uses the right approach.

“It has to be done carefully. You can’t go in there saying, ‘I’m smarter than you,’” Okabayashi said. “It’s never going to go well. My experience with scouts is that they’re really friendly and easy to talk to. I found that surprising.”

The more he dove into that world and had those conversations, the more his frustrations with the current data in hockey analytics started to grow. A scout would make an observation about a player that he would see as well -- an attribute that was important to winning -- and he couldn’t always quantify it.

The data the Devils were collecting and manipulating wasn’t all that much different from what we see publicly. The big analytics firms like Stathletes have the bandwidth and power to go deeper than small analytics groups within the teams.

For instance, Stathletes produced an in-depth 20-page report on P.K. Subban vs. Shea Weber before the Canadiens pulled off that trade. The data in that report is almost overwhelming with breakdowns that include zone entries, expected goals per 100 passes, success rate of zone entries allowed on penalty kills, zone clears, play-to-advance ratio, location of puck touches, type and quantities of puck recoveries, defensive zone battles and tons more.

The data Okabayashi worked with in New Jersey was similar to what public hockey analytics experts examine.

“It’s no secret; I think everybody cares about possession and scoring chances. And really, we spend a lot of time defining the nuances around that,” he said. “We got to work with the video staff to define other metrics we thought were important and deliver those to the coaches as well. That’s just going to be an ongoing evolution.”

In working with the Devils, Okabayashi got a taste of what’s available right now in the world of hockey analytics. With it came the realization that there is still a long way to go because the data being produced right now still isn’t good enough to be the only source of information on which to make decisions.

For someone dealing with numbers, that can be frustrating.

“Yes, we have some metrics, and we provide some useful information, but not as concrete and touchable as I’d like [the stats] to be, considering what you get watching games and talking to scouts,” he said. “My friends who work in basketball, they’ve identified the pick-and-roll as the point-generating play. Some absurd amount of points come straight from the pick-and-roll play. Teams dissect the pick-and-roll. There really isn’t -- maybe with the exception of special teams -- you don’t have the same in hockey. It means the analytics around the sport are more challenging. It’s a long way of saying if I thought the analytics as they exist today were more powerful, I would be pounding on people’s doors saying this is absolutely the right metric.”