Revenge of the Nerds: NHL.com embraces advanced analytics
Ready or not, it's time to embrace advanced stats.
With a flick of a switch on Friday, the NHL kicked off a new information age, offering a vast array of new statistical tools to the public that bring modern analysis out of the basement and into the mainstream.
At least, that's the theory. It remains to be seen whether ready access will lead to greater acceptance of these stats among the masses, but by embracing these concepts, the league has vindicated the work of the stat geek pioneers and opened the door for some amazing projects that are just around the corner.
For now, though, nothing too fancy. Listed on its official web site under a tab labeled Enhanced Stats, the NHL is offering five season's worth of data captured in 30 new categories. All of them are built on models that have been in use for years by the hardcore stats community, a small but fanatical group of fans who scraped and molded existing data to quantify various elements of the game.
The holy grail among the faithful has always been a number that accurately captures puck possession. While that formula has remained elusive, there's been widespread acceptance for shot attempt-based proxies known as Corsi and Fenwick. The problem with those is that the nomenclature is impenetrable to the casual fan, so the league ditched those names for the simpler SAT (shot attempts—shots on goal+missed shots+blocked shots) and USAT (unblocked shot attempts). Those will be broken down into attempts for and against while a player is on the ice and then broken down again to reflect score situations—even, behind and ahead.
The league will also report PDO, which looks at a team's shooting percentage and save percentage numbers to suggest its level of puck luck. That stat will be known as SPSV% (shooting percentage, plus save percentage).
Depending on how far down into the rabbit hole fans want to travel, they can use these numbers to try to quantify what a player brings to the table. For example, Patrice Bergeron leads all forwards with an SAT differential of 118, meaning that when he is on the ice the Bruins have attempted 118 more shots than they have allowed. That seems like validation for his reputation as the top defensive forward in the game.
At the other end of the spectrum, Buffalo's Rasmus Ristolainen ranks last with a -200 differential...which is probably what you'd expect of a rookie defenseman playing on the league's worst team.
But like any other stat, these are far from perfect. A player can get credit for or against whether he's in the middle of the play or heading off for a change, whether he's the one making the attempt or not, or whether he's covering his defensive assignment or counting butterflies two zones behind the play. But taken in context over a large sample, the numbers are useful and sometimes even predictive.
And the numbers go well beyond proxies into harder data, including zone starts, quality of competition, penalties drawn, points per minutes played and shot distance, and it even includes assists broken down into primary and secondary and scorers sorted by height, weight and age.
Taken together the league's plan offers not just numbers, but greater context than ever. And this is just the start. The league's long-term plan calls for a four stage rollout.
Phase 2 will include more enhanced stats options with a focus on increased visuals—charts and graphs that will allow users to compare players or teams over various periods.
Phase 3 will feature more data visualization and new data mining in an effort to create more useful stats. It could also, depending on settling issues with the NHLPA, involve player tracking technology that could provide that holy grail of possession data, possibly as soon as the start of the 2015-16 season. This could possibly be hidden behind a paywall, or used in conjunction with another product like NHL GameCenter.
Phase 4 will bring a complete digitized record of every hard copy game sheet from throughout the history of the league. It's also expected to include access to every goal that has been captured on video.
It's an ambitious plan, and whether you're an advanced stats apostle or just dipping your toes into the shallow end, everyone can agree that more information is good. Now that it's out there, it up to fans to decide how much or how little of it is of use to them.