-
Greg WyshynskiESPN
SAN JOSE, Calif. — John Tavares and his Atlantic Division All-Star teammate Steven Stamkos skated into the offensive zone. All three players from the Metropolitan Division were bunched up around Tavares, leaving Stamkos wide open for a goal.
A fan watching his first hockey game could have diagnosed this defensive breakdown — egregious, even by 3-on-3 exhibition game standards. But on NBC’s augmented broadcast, featuring player and puck tracking data collected by sensors embedded in shoulder pads and inside the puck itself, fans saw the rest of the story: a blue CGI triangle linked the defenders, identifying how many feet away they were from the puck; and statistical information flooded the rest of the screen, from traditional stats to a running count of the players’ ice time. At other times in the broadcast, a subtle gray streak trailed the puck in order to make it easier to follow on a smaller second-screen experience.
It was, at times, information overload.
But by next season, it’ll be the new normal.
Puck and player tracking has arrived, although like any emerging technology, it’s still being tweaked and revised. It will be implemented beginning in the 2019-20 regular season, changing the way we watch the NHL, how broadcasters analyze games, how teams evaluate their players and, soon, how gamblers wager on hockey.
“We’re committed to ensuring that the NHL is at the forefront of all new technology,” commissioner Gary Bettman said prior to this grand test of the tracking system at the All-Star Game at the SAP Center. “We think many of our fans — especially the innovation generations, the millennials and Gen Z — are going to love this new frontier.”
And the players?
“More analytics for the GMs to hold against you,” Columbus Blue Jackets forward Cam Atkinson said with a laugh. “No, I think it’s good for the fans. It looks pretty cool, to be able to track your favorite player when he’s on or off the ice.”
#HiTechHockey pic.twitter.com/PlrOrabErt
– NHL on NBC (@NHLonNBCSports) January 26, 2019
At this point, there are more questions than answers about puck and player tracking, from its usefulness to its role in player evaluation to its potential for a gambling windfall for the NHL.
One question that can be answered with some certainty: How did we get here?
From glow puck to freezers
The experiment began in 1996 with FoxTrax, forever infamously known to hockey fans as the “glow puck.”
It was a failure, to be sure: a garish blue haze that could barely keep up with the puck on the broadcast, followed by a cartoonish red comet tail that could’ve easily been confused with the graphics of some off-brand video game. But just as 3M engineer Spencer Silver’s glue that didn’t really work would eventually result in the Post-It Note, the tech behind the glow puck would be used practically in NASCAR (identifying drivers around the laps) and in the NFL, in the form of the first-down line. Its array of infrared emitters and the electronics placed inside the puck would inspire the NHL’s puck and player tracking nearly 20 years later.
In 2014, the league began to dabble with it again, thanks to Bettman’s desire to enhance the television experience. “He felt that we had been presenting the game the same way for a long time. One camera, panning back and forth. We have 60 minutes of actual play. Compare that to the NFL, when they have a lot of time to feed you graphics and stats. It’s great. But we don’t have that luxury. We have to do it in real time,” said Dave Lehanski, senior vice president of business development and global partnerships for the NHL.
The NHL partnered with Sportvision, who coincidentally was responsible for that CGI first-down marker in the NFL. At the 2015 NHL All-Star Game in Columbus, a series of cameras was placed around the arena that tracked not only the players, who wore special tags on their sweaters that emitted a signal, but the puck itself, which had infrared sensors built into it. The data this tech collected was revolutionary: from speed and ice time to geographic information that could digitally recreate the action in real time or, later, in a virtual reality environment. The experiment worked well enough that it was tested again at the World Cup of Hockey in 2016.
There were challenges with the technology. For example, not all arenas were built the same way, so players in the penalty box would sometimes be counted as still being on the ice, depending on the placement of the cameras. But the biggest issue was cost, not only to install the technology in the arenas but within the puck itself.
“It actually worked well from an accuracy standpoint. But no one really liked the look of the puck,” said Lehanski, “and there were concerns about the way they were manufactured. We took existing pucks and cut them in half, hollowed them out and put the stuff inside.”
Early estimates were that these enhanced pucks — which featured exposed infrared emitters on their sides — could cost in the range of $100-$200 apiece, although Sportvision would later claim those estimates were overstated. Still, it harkened back to the early days of the glow puck, when then-NHL senior VP of hockey operations Brian Burke said the league lost $1,200 just from pucks flying into the stands during its debut game.
The NHL sought to reduce cost by going to an optical camera-based tracking system that was favored by sports like football and basketball. One problem: “The puck moves really fast,” said Lehanski. “It’s constantly changing, and moving behind other objects — skates and sticks — which makes optical obsolete.”
So after that flirtation, it was back to the “chip in the puck” idea, this time with a focus on ultra-wideband radio frequencies being used to track wearable technology on the players and inside the puck. “There were several companies that we met with that said they could track the puck. We were like, ‘That’s great … now tell us how you manufacture it,'” said Lehanski.
That’s how the NHL ended up with Germany’s Fraunhofer Institute as its player and puck tracking partner. They built a puck that was within a “tight range” of the actual puck’s weight. They built a system that no one else had: a freezer that stores the puck, charges it, tests its battery life and signal strength. They built a machine that has a mechanical arm with a syringe that pumps an ideal amount of adhesive to seal it, and then applies 10 tons of pressure.
Jogmo World Corp. partnered with Fraunhofer to create the technology used in player and puck tracking. There are 14 antennas placed around the arena. There’s a sensor in the shoulder pads being tracked 200 times per second, and a sensor inside the puck that can be tracked 2,000 times per second. It was tested at the 2018 and 2019 NHL All-Star Games, as well as at New Jersey’s arena and during two regular-season games at the Vegas Golden Knights‘ T-Mobile Arena — timed to sync up with the Consumer Electronics Show for maximum impact.
Lehanski said these tests reveal a level of trust between the NHL and its players.
“We’ve made huge strides with them. A year ago, I’m not sure I would have thought it was possible that we could do what we did in Las Vegas, putting our tech on players in a regular-season hockey game,” he said.
But that trust only goes so far.
The players vs. tracking
Jack Eichel is wary of the new pucks.
“I hope it doesn’t change the puck at all. I feel like maybe with a tracker in the puck, it could change the feel of it,” said the Buffalo Sabres star. “You’d be surprised how close we pay attention to that stuff. The gear isn’t as important as the puck. Guys can feel it.”
This is a problem that tracks back to the glow puck. An official puck has to be between 5.5 and 6.0 ounces, with official pucks usually clocking in right at 6. The glow puck, when it was first tested, came in at about 5.9 ounces. “There were players that just didn’t like the puck. They claimed that it had more rebound. And perhaps it did. If so, it was small, but they’re professional players and perhaps they could tell,” chief engineer Rick Cavallaro of Sportvision said in a recollection of FoxTrax.
There were some players in the Vegas experiments that felt something off about the pucks. But Lehanski isn’t concerned. “It was expected. It was a regular-season game and we think the puck was in line with the specs. But it was still a test. It’s an evolution. We want the feedback. It’s not a bad thing that players are telling us that they want a better puck,” he said.
(Internally, the NHL expects the same life cycle for the chipped pucks as it does for any major change in rules or equipment: ride the criticism through until December and then probably never hear about it again.)
Most players ESPN spoke to about player tracking were more concerned about what the sensors are collecting than the sensors themselves. It’s not just how these speed and stamina stats could expose certain players, but how they could be used by any number of parties.
So the NHLPA and the NHL came to an understanding on some subjects, like where the raw data will end up. There isn’t a formal agreement yet, but it’s their understanding that the data is going to be proprietary for the teams and the broadcasters, and parceled out on NHL.com, perhaps in the form of some leaderboards. But fans (and stat-heads) looking for huge data dumps of tracking info from which to glean better possession numbers or create new stats will be disappointed — at least at the start.
“I think it was fair to say that there were concerns in the beginning. They were super curious and had a ton of questions. So we had to explain the system, what you’re going to get out of it. And then they asked about how the data was going to be used. So no, the raw data isn’t going to be out there in the beginning because they need to be comfortable with what we’re tracking,” said Lehanski.
“We haven’t come to a formal agreement about that yet among the players and the clubs. But me, personally? Yeah, I could see a day in the not-too-far future when we’re putting some — never all — of the raw data out there, and I think it would be great to crowdsource new thinking. Like, what could be our version of a quarterback rating for goalies?”
Much more formal is the written agreement collectively bargained by the NHLPA and the NHL regarding the use of tracking data in contract negotiations, putting in “protections” for the players that will last until the next CBA is ratified.
“Typically, the clubs aren’t supposed to rely on any of the data we’re collecting in player and puck tracking in contract negotiations,” said deputy commissioner Bill Daly. “It’s not just salary arbitration. It’s contract negotiations generally. Citing speed [or other data] and relying on those statistics in negotiations.”
Daly said that some players were concerned that contract talks could become too “stats-reliant” with this data. “Some players that don’t have those attributes, but have other attributes, are concerned about that,” he said.
But that agreement cuts both ways: Players won’t be able to use tracking data to bolster their contract cases, either. Which struck many observers as odd, because the more data that’s on the players’ side in a negotiation, the more arguments they can make about their value beyond traditional stats. This has been a hallmark of the analytics era, but the players made sure it couldn’t be used in the tracking era.
“Sometimes you hear from people who would say, ‘Look, I know what my job is, and if I’m a half second slower than I was two years ago, as long as I’m there when I’m supposed to be, what’s the difference?'” NHLPA executive director Donald Fehr told ESPN. “The critical thing here is, we didn’t have agreement on everything, but we got agreements officially to do the experiments and see how it works. Then we’ll see what the data is, we’ll see if it’s helpful in other licensing agreements, we’ll see if the broadcasters like it, and we’ll see the rest of it. Then players, along with their agents and everybody else, can look at it and say in that particular respect, does this make any sense? Should it be tweaked? Should it be reversed? This is short term. This is baby steps.”
Mathieu Schneider, special assistant to the executive director of the NHLPA, said on Friday that there was a split among the players on acceptance of the tracking technology.
“In a sense, you have a bit of a clash between some of the younger players and some of the older players. The first thing is we should have more accurate stats, which is a big part of this that goes under the radar. The second thing is that it’s going to enhance the broadcast and tell more stories about how great the athletes are in the NHL,” he said. “The positives are going to outweigh the negatives when this all shakes out.”
Betting on player tracking
The inspiration for puck and player tracking wasn’t for wagering purposes. Then the world changed.
“Before Columbus, no one was thinking about this. There was no foresight into the laws changing anytime soon. And then two years ago, all of a sudden that landscape changed,” said Lehanski, on the march of legalized sports wagering in North America. “We started to think about the data and how that could provide value to the betting community. That just bolstered the strategy.”
It also did something that couldn’t be accomplished four years ago for puck tracking: it justified the investment. The NHL is cutting licensing deals with sportsbooks — including MGM Grand, its first announced partner — for its player tracking data. It’s not only something that will help them better establish their odds, but it’ll open the doors to legions of prop bets during games.
“We’re not going to pursue something at a significant cost if [Bettman] doesn’t know we’re going to be able to make it back in the multiple,” said Lehanski.
Daly said that the NHL and NHLPA have a licensing agreement in which player names can be used for wagering purposes, such as prop bets based on the tracking data. The licensing of that data is considered hockey-related revenue, but the players don’t get anything additional for having their names used with wagers. Fehr felt that could change. “Nobody really knows. You have to try it and see what it does, what revenue it produces, and see if it has any effects going the other way. All of this is baby steps and experimental, I think,” he said.
The wagering potential for puck and player tracking is incredible. During the tests in Vegas, prototypes of mobile wagering were displayed that would allow fans to make in-game bets that included:
-
The portion of the ice from which the next goal would be scored. In other words, you can get odds on Alex Ovechkin scoring from the “Ovi Spot” on an upcoming Capitals power play.
-
The portion of the net the next goal will be scored into. High glove side was a minus-230, while the five-hole was just minus-180.
-
Betting an over/under on the velocity of the next goal.
-
Which team will possess the puck longer, with a running clock on their possession times.
-
The ability to wager on two players and which one will skate the farthest distance in the game.
-
The ability to wager on which goalie will have saved the faster shots on average in a game.
“Our original goal in creating this technology was to create a broadcast enhancement that can be used in real time, which no other sport has. While not developed with this in mind, the application to sports gaming could create even more fans,” said Bettman.
But will it? That’s the question. On background, there’s skepticism from many parties — including in the sportsbook community — that hockey generates enough betting action to sustain prop bets. But there’s also an optimism that this is uncharted territory; and that at the end of the day, it could be something to attract new fans to the game, and to the arena.
“In the beginning, it’ll be about marketing. But no one really knows how many people want to wager on how far a player skates in a game,” said Lehanski. “Once you’re in the game, all of that stuff looks attractive.”
Who explains this stuff?
Think back to that scene from the All-Star Game, with John Tavares surrounded by three defenders. Through player tracking, we knew how far away from the puck they were, through that little triangle on the ice.
What we didn’t hear is why that matters. How far away should they have been? What does the exact distance of that spacing mean?
Extrapolate that to other events defined by puck and player traffic. We know how fast a player is going, but what does that mean vs. his own history or the speed of his peers? We know how fast the puck is traveling, but what does that mean with regard to where it is and who is shooting it?
The “why” of things is going to be one of the biggest challenges in the tracking era, especially in the media. There needs to be another level of analysis — especially on broadcasts — that takes this tracking data from trivial to vital. Visualizations, charts and graphs are great; explaining what they mean is a more difficult, and much more meaningful, task.
This is where the players are worried, not only in public consumption of the data but in how teams might use it.
“I think to a certain extent, you might be looking for the right things in the wrong places. I’ve seen guys that skate fast and shoot hard that can’t play a lick. So I think there’s some usefulness to it. But if a pitcher throws at 110 mph but can’t throw strikes, he’s no good for us,” said Winnipeg Jets forward Blake Wheeler. “I think smart organizations … it might be a piece of the puzzle. But if you’re basing your signings on just a couple of things, that’s probably not the best way to go about it.”
The players’ mantra is the same we heard during the rise of analytics: that there are intangibles that go beyond the collected data. “There are a lot more to hockey players than just the stats. When you look at Brett Hull, Nicklas Lidstrom and Wayne Gretzky, they were never defined by how big they were or how fast they were or how hard they shot a puck,” said Schneider.
But there is a segment of the NHL population that has been craving this added context: goaltenders.
“We’ve been in the goals-against average and save percentage era for about 100 years, and that’s it. Now, we’ll know where the shot was taken, who took it, was it deflected, were there players in front of goalies. Fans can come up with their evaluation of goalies, and NBC and Rogers can come up with theirs,” said Lehanski.
“On the broadcast you’ll hear them say a goalie is on his game, that he’s playing big in the net. Non-hockey fans don’t know what that means. But now we’ll show you in visualization the distance he’s coming out from the net, whether that’s more or less than usual, whether he’s on his angles. We’re doing this to call out the players’ skills and speed, but also educate the fans.”
Dallas Stars goalie Ben Bishop has been a critical voice in the past about goalie analytics. “There’s so much more that goes into a read for a goalie: Who’s shooting it, from where, who’s in front of you, who’s backdoor, what are his outlets? So many more variables than who scored a goal and from where,” Bishop told ESPN.
There’s a chance with puck and player tracking that these nuances could be better established, especially when it comes to the digital recreation of players thanks to the data collected by it. Jon Waldman, the NHL’s director of strategic planning, tells a story about a soccer game that used tracking technology. A free kick appeared to freeze the goalie, leading to criticism from the broadcasters. A digital recreation of the play, rotated to the goalie’s POV, revealed subtle shifts in the wall of players in front of him that exonerated him.
“It’s such a tricky thing for goalie stats,” said Minnesota Wild goalie Devan Dubnyk, a proponent of player tracking.
Like some of his peers, Dubnyk is a proponent … as long as they’re not tracking his play.
“My secret’s going to be out that I don’t go anywhere. There’s just going to be a black dot in the middle of the crease, and that’s going to be my total distance traveled,” he said, with a laugh.
Of course, there might eventually be another application to this technology that will impact Dubnyk and his peers. We have a chip inside the puck. Could we finally have a system that can determine if it crossed the goal line, beyond the current one of obstructed camera angles and blurry replays?
“The potential is there,” said Daly. “I don’t think it’s our first utilization, but it’s a possible utilization at some point and time.”
Technology can be re-purposed to solve adjacent problems. Hey, no one thought the glow puck would lead to accurate stats and a revolutionary evolution of the way we watch and bet on hockey. Yet here we are.