Shortly after speaking to a crowd of 300 people, Trey Elder was a popular guy.
Elder was one of four presenters during the HALO conference and meetings Tuesday at Ball Arena. Hosted by Arik Parnass, who leads the Colorado Avalanche analytics department, the conference was the largest collection of NHL analytics staffers the league has ever seen. Each of the presentations was a finalist for the conference’s hackathon competition.
A graduate student at the University of Pennsylvania, Elder gave a presentation on a metric to better define forechecking success called PRESS (puck recovery and exit suppression score). Shortly after Elder’s presentation, there were several attendees who wanted to speak further with him on the topic.
“It’s been great,” Elder said. “I’ve never presented at a conference before, let alone in front of like, real NHL people. It was very cool. I was a little nervous, like I think everyone would be, but I felt very prepared, so I feel like I did an OK job.
“I tend to think that no one is going to solve hockey, or come up with a singular thing that is like, ‘This is how you win more games,’ or ‘this is how you score more goals.’ I’m very interested in the aspects of the game that are somethings like the game within the game. With forechecking, the object isn’t to score a goal, it’s to get the puck back. That’s an aspect of the game that feels like it could be optimized in some way.”
Elder grew up in Northern Virginia, rooting for the Washington Capitals. His goal after school is to some day work for an NHL team.
As part of the hackathon competition, all of the contestants were given the same set of data — a batch of AHL games from two seasons ago — to work with. Elder and his partner, Jonathan Pipping, went through about 28,000 forechecks, and developed a model for league-average success rates in two components of forechecking.
They were then able to score players against that average. One member of an NHL team’s analytics staff said the findings from Elder and Pipping lined up pretty closely with what his club’s research into forechecking success has found.
Colorado Avalanche forward Joel Kiviranta didn’t spend much time with the Colorado Eagles in the AHL, but he was there during the timeframe of this set of games. He also scored as one of the top-five forecheckers in Elder’s metric.
“I’m sure there are groups that are collecting that data and analyzing it however they want to analyze it,” Avs forward Logan O’Connor said. “As the game evolves, analytics have obviously become a huge part of sports in general, and especially with hockey. It would definitely be fascinating to see the tendencies and player-to-player matchups, how teams do it differently, and what the success is, where the puck is, where the player is, and whatnot.
“I think at some point analytics can become overwhelming, because hockey is so dynamic. Baseball analytics are a very easy, set number of things that they look at, whereas hockey is so fluid that I think it can definitely get challenging. I’m sure those guys are brilliant for a reason. I’m sure they have ways around it.”
The Avs are one of the most aggressive forechecking teams in the NHL. It’s the the fuel for their offensive engine. Colorado wants the game played at a frenetic pace, and dialing up the pressure far away from their own goaltender when the Avs don’t have the puck often allows them to create turnovers and continued offensive threats.
NHL teams have been trying to identify the best forecheckers for as long as there has been scouting. Avs coach Jared Bednar has said how his team forechecks is one of the first indicators he looks at to see if his team came ready to work in a particular game.
“I mean, I would love it,” Parker Kelly said of a standardized forechecking metic. “As a forechecking guy, yeah. I feel like there’s got to be forechecking stats out there already. Maybe it’s not a league-wide set, but we have our own in here and that’s a big stat we do look at.
“I feel like that’s kind of a staple of our game. When our forecheck is humming, it’s usually a good indicator of our legs. We’re hunting the puck. We’re making good decisions. I’d love to sit down and go through all of those types of stats. I find them really interesting.”
It takes hard work to forecheck, but there’s more to it than a player just skating around and throwing his body around to be an effective forechecker at the highest levels. There is skill and craft involved. Most of the best forecheckers don’t rely on big hits.
Guys like Kelly, O’Connor and Kiviranta are well-versed in the other team’s plan to get the puck out of danger. There are times when the best forechecks don’t involve hitting at all, but rushing a defenseman into making a poor decision.
As Elder noted in his presentation, the forechecker who creates the original turnover often never touches the puck and therefore never gets any credit in a traditional box score. Elder’s model hopes to define better credit for players at each level of a forechecking system.
One of the attendees at the HALO conference was Andrew Cogliano, special assistant to Avalanche general manager Chris MacFarland. Cogliano was one of the league’s most respected forecheckers for years. It’s not hard to imagine a world where, with more defined and readily available data like Elder’s PRESS metic on who the best forecheckers are, a player like Cogliano being an even more valuable commodity.
“I think with the skill of it, in regard to things like body position, angling, stick position, that’s something that would be hard to put a metric on,” O’Connor said. “It is a bit more challenging, but as you average things out over a season or a career, you could definitely see tendencies and who specializes in it, and who can make a career out of doing that, for sure.”
Want more Avalanche news? Sign up for the Avalanche Insider to get all our NHL analysis.