Alain Vigneault’s Internal Metrics Hold No Weight
The claim is not a new one, even if we don’t hear it very often.
When things aren’t going well for the Rangers and questions about advanced statistics make their way back to Alain Vigneault, he reminds everyone of it: he has his own internal metrics (sophisticated stats) that he uses to make decisions.
A few weeks ago Steve Zipay wrote a blurb based off Vigneault’s comments on the matter:
A popular stat such as Corsi, for instance, doesn’t measure speed or determination or chemistry. It measures possession: shots on goal, missed shots on goal, and shot attempts blocked toward the opposition’s net at even strength while a player is on the ice, minus the same shot attempts directed at your team’s net. McDonagh has been deep in double-digit minus territory in some games.
Indeed, the Los Angeles Kings led the league in Corsi for five years, including the Stanley Cup championship seasons of 2012 and 2014, but began to struggle and missed the playoffs two of the past three years. Shot attempts weren’t being converted into prime scoring chances or goals.
Neither Vigneault nor McDonagh are overly concerned about Corsi values.
“I’m not a big shot-attempt preacher or believer,” Vigneault said. “We have one of the most sophisticated stats packages, that has evolved through the years, that gives us breakdowns of important things we believe in. I know the shot attempts, Corsi . . . I’ve got other things I rely on. I don’t put too much emphasis on that.”
Zipay’s continued attacks on advanced stats is both laughable and expected, so we can ignore that. It’s more Vigneault’s comments that should raise your eyebrows.
Does his “sophisticated stats package” contain baseline advanced stats you might find on, say, Corsica? Doesn’t seem like it. Something you might be able to track yourself? Maybe, if you were able to see it. Something that translates to winning or positive play? Not by any accepted public metric.
Before we can go there, though, we need to start from the beginning. Vigneault has always maintained that he believes in statistics, only he does so with a catch: he believes his own statistics. Vigneault and his team track and compile statistics after every game. If you would remember, after a playoff game a few years ago Vigneault cited the team creating good scoring chances in the contest; it’s worth noting by traditional metrics they were not.
Outside of those rare moments, though, Vigneault talking about the statistics is as mysterious as the statistics themselves. We don’t know what they are or what they measure – or even, you know, how they’re measured. It’s worth noting that certain statistics are somewhat subjective, specifically a stat like scoring chances, where the totals might differ depending on who is making the call on what is and isn’t a chance. That being said, when the data is publicly released for evaluation, issues can be fleshed out and analyzed. Over time, when information is public, a norm eventually gets set; as you’ve inevitably seen from places like Corsica or almost every credible stats-provider in the field.
When information is not available for public consumption it becomes immediately suspicious. I understand why Vigneault wouldn’t want to share his data, especially if he thinks it’s ahead of the curve, since coaches are looking for any edge they can find, but when the product on the ice doesn’t reflect success then what exactly are we looking at?
For a coach who doesn’t make major changes, this is the biggest and loudest of the red flags. For years Dan Girardi manned the top pair despite being unfit for the job. For years Tanner Glass played over better, more qualified players in the bottom six. Last year the Rangers blew three separate leads with less than five minutes remaining (and two with less than two minutes remaining) by relegating Brady Skjei - Brendan Smith (who were spectacular in terms of public advanced metrics) for Marc Staal - Nick Holden (who were at the bottom in terms of public advanced metrics) in all three instances (and others!).
Despite the Rangers offloading one of the biggest hauls in recent memory for Keith Yandle, the superstar defenseman was used like a third-pairing guy and got less power play time on Broadway than ever in his career. And in case you thought that lesson was learned, Kevin Shattenkirk has seen as much time on the third pair as he has on the top pair this year.
Recently, the adjustments made to the lineup have been just as concerning. Vigneault broke up the Buchnevich - Mika Zibanejad - Chris Kreider line for no reason, despite the trio being the only source of consistent offense on the team. He moved Buchnevich to the fourth line, then to the third, and doesn’t give him ice time for no discernible reason. The power play has been drowning the past month and yet guys like Kevin Hayes can’t get a shake on the second unit.
Time is very much a flat circle with Vigneault behind the bench. Look no further than this article from his firing in Vancouver and tell me that you can’t just replace “Vancouver” with “New York” and “Canucks” with “Rangers” and have it be just as accurate. Here’s just a few bullet points from that story (emphasis mine):
During his time here, Vigneault displayed an inability to consistently nurture young talent. In the salary cap constrained world, getting contributions from young players with small salaries is paramount.
Clearly, young players who were committed to defense first (like Tanev and Hansen) would be in the good books of Vigneault, while those with creative offensive instincts (Hodgson and Kassian) would be forever shackled.
Since the Stanley Cup loss to Boston two years ago, the number of sixty minute efforts expended by his squad could be counted on one hand. The President’s Trophy and divisional titles in that span were buoyed by a weak schedule and strong goaltending.
Circling back to his statistics; I don’t think it’s a stretch to be able to state that his stats aren’t good at evaluating things that lead to success. While we don’t actually know what the stats are, the product on the ice has gotten progressively worse under his guidance. The 2014 Stanley Cup run was the result of Vigneault just coaching, with absolutely no input on personnel decisions and without him being in tune enough in the locker room to find “his guys.” The moment he got his hands on the roster that summer Anton Stralman, Brian Boyle, and Benoit Pouliot (this was defensible at least) walked out the door for Dan Boyle and Tanner Glass. Carl Hagelin was traded for a single NHL asset (Emerson Etem); he wasn’t used and eventually discarded for an asset currently playing in the KHL.
Again, if you thought the lesson was learned, Derek Stepan has yielded not a single NHL return as of this writing (with Anthony DeAngelo back in the AHL). And the king’s ransom acquired for Matt Duchene only makes it look worse – but that’s a Jeff Gorton issue.
Before the Rangers got shellacked by the Pittsburgh Penguins in 2016, we warned things weren’t as good as the media was making it out to be (here, here, and here are just a few examples). Last year we gave similar warnings, and then continued to yell and scream about the defensive usage as it sunk the season and the series (here, here, and here). The media plays a role in this in terms of the public eye, as they’re quick to defend Vigneault (who is a dream to work with compared to John Tortorella) based on the simplest of qualifications.
"NYR are a playoff team right now. Maybe AV knows what he's doing, eh bloggers?" - Xmas 2015— HockeyStatMiner (@HockeyStatMiner) December 27, 2017
"NYR are a playoff team right now. Maybe AV knows what he's doing, eh bloggers?" - Xmas 2016
"NYR are a playoff team right now. Maybe AV knows what he's doing, eh bloggers?" - Now
Some lessons are never learned.
Again, back to statistics. The entire point of using these units of measurement are not to perfectly forecast success, since that’s damn near impossible, but rather to at least get a gauge of what is and isn’t working. Statistics aren’t everything, but they give color to what it is you’re seeing. So many things happen in a given hockey game and they happen to every single player. It’s impossible to simply use the eye test to make accurate, lasting impressions on everything you’ve seen; and since the brain tends to shift towards recency bias or extreme events, it makes it even harder.
Stats do have a place in sports, whether you like it or not. At this point, ignore them at your own risk. The Rangers aren’t doing that, though. They are using stats, just their own, which is almost more dangerous. If a team like the New England Patriots or the Pittsburgh Penguins want to claim they use their own stats -- after backing it up with championships – that’s one thing, but the Rangers haven’t been nearly good enough to make this claim. Instead, it’s become something of a get out of jail free card: media members who don’t have teeth anyway use it as a defensive mechanism, and Vigneault gets to walk about without having to answer to the failures.
On the podcast this week, I talked about how the media’s reporting shouldn’t have a direct impact on the things that are happening on the ice – at least not right away. But constant reporting on problems and things the fans seem to be interested in, do eventually force a coach to take a harder look at what it is they’re doing. Remember that when Kreider scored the game-winner in overtime of Game 4 against Boston in 2013, Tortorella – who was routinely roasted by the media for not playing him more – told the media after the game he was happy about the kid scoring because the media had been killing him about it. Funny how when a coach comes with a little more honey, a similar situation (Buchnevich) is largely ignored.
Regardless, I think it’s fair to say the Rangers sophisticated stats package isn’t good enough. And the fact that the team continues to put stock in what it tells them isn’t helping matters either.
Because the best thing about numbers? They don’t lie. Unless you position them to tell you and track what you want to see rather than what the truth is.