It’s the little secret that doesn’t get mentioned nearly enough when it comes to the “advanced stats” that seem to be the hot topic in hockey circles right now.
Those stats? They’re not very advanced at all.
Measuring which teams and players play in the offensive zone more than in the defensive zone, using a percentage, is what those stats often come down to.
That’s not to say that numbers like Corsi, Fenwick, zone starts and the rest don’t have their uses. They can be, in fact, powerful tools when it comes to evaluating teams and individual players in ways that the old stats – goals, assists, points, etc. – couldn’t.
More than any other off-season, members of NHL management teams caught on to that this summer. Many started using these possession numbers to complement the information provided by scouts.
They also hired some of the bright lights (Tyler Dellow, Eric Tulsky et al) who had been analyzing that data online to help with that process.
“We’ve taken some of the analytics outliers – players with very high or low Relative Corsi – among free agents and then have gone back to video and tried to see if we agree with the numbers,” one NHL executive explained recently, relating how his team used the numbers produced by sites such as extraskater.com, whose founder, Darryl Metcalf, has since been hired by the Toronto Maple Leafs. “On a couple of them, it’s been very illuminating.”
But current data is pretty basic compared to what’s coming. Hockey’s analytics are about to take a major leap, one that will begin to intertwine data and video using new technology already being employed in other sports.
As reported by Marc Antoine Godin in La Presse on the weekend, the NHL is investigating a system that would put a small chip or RFID tag on its players as a new way to track their movements and collect information during games.
This, along with puck-tracking, would provide teams with the precise speed of players, their shots and passes, and even more elaborate data, such as which players enter the zone with possession of the puck the most, and which ones recover it the quickest in their own end.
The possibilities would theoretically become almost endless given how much data this kind of tracking can provide beyond what’s currently available.
And it would be given to every team to use as they saw fit.
“I think there would be some standardized sharing of the statistics,” the NHL’s chief operating officer John Collins told Godin. “Then clubs would be free to use it, slice it and dice it anyway they think of, based on their own philosophy on how to integrate statistical data into coaching, training and everything else.”
At the earliest, it’s expected the NHL would go this route by the start of the 2015-16 season, provided the NHLPA agrees to allow players to be tracked.
In the interim, some teams have been pushing to find an edge on their own, hoping to get out in front of the competition. At least one is believed to be using this kind of tracking technology in practices. Several others are investigating buying into new options to track games.
One company, PowerScout Hockey, has been working with several clubs using a demo from several games last year in the hopes they sign on for this season. At least a half-dozen are believed to be seriously considering doing so despite a rumoured price tag of nearly $100,000 a season. (Some teams rationalize that cost by comparing it to the salary of a pro scout.)
Another firm, SportVU, is more closely aligned with the NHL and is widely expected to be the group that will eventually implement the kind of system Collins is talking about. SportVU’s advantage is that it has an agreement with the NBA, which potentially offers easier installation and access for the nine NHL teams that share an arena with an NBA team.
Both PowerScout and SportVU could initially use camera-based systems to track player movement without chips or tags, but that process is only partially automated and can require a 24-hour turnaround to produce data for even a single game.
One problem with camera-tracking that the NBA has already run into is the huge volume of data that’s produced. Sorting out what’s useful information and what isn’t could ultimately be a process that takes NHL teams – and their recently hired analysts – several years.
At that point, hockey will be at the advanced-stats stage, with the concepts and calculations moving increasingly behind closed doors and beyond the average fan – the way they have in baseball.
That’s part of what makes this coming season an interesting crossroads for the NHL. Some teams are deep into this type of analysis already, investing big money, while others are just starting to take that leap.
A third group of clubs aren’t yet on board and remain skeptical that what’s available can help.
The only area where there’s beginning to be some consensus is with simple possession data, as teams such as the Edmonton Oilers and Pittsburgh Penguins made player acquisitions this summer aimed at improving in that area.
If you look closely at some of the signings made in free agency, it’s evident that puck possession – as measured primarily by shot attempts at even strength – is progressing from an undervalued stat to one much more commonly used by general managers and coaches.
At some point, it could even shift to overvalued, as players who drive Corsi continue to get paid for that trait alone.
For now, however, there’s an advantage in understanding that relatively simple data and its concepts, primarily because some teams aren’t taking the time to deep dive into it.
With RFID tags and camera tracking on the way, that advantage likely won’t last much longer.Report Typo/Error