Advanced stats vs. the old guard: inside the bitter rivalry
Advanced stats vs. the old guard: inside the bitter rivalry
Why do the stat geeks and dinosaurs hate each other so much? And why does the advanced stat debate have to be so black and white? THN takes you inside the rivalry
It’s late June 2014. Tyler Dellow and Steve Simmons want to kill each other. Hyperbole? Probably. At the very least, though, Simmons is about to boil over and Dellow calmly relishes it.
Dellow, a Toronto-based lawyer, and Simmons, a Toronto-based sports columnist, are debating the validity of advanced statistics on a radio show and, more specifically, statistical darling Mikhail Grabovski versus clutch playoff performer Dave Bolland. Simmons tells Dellow, one of the leading voices in the analytics community, to throw his stats out the window and look at Grabovski’s lone-wolf tendencies as a center.
“I just judge hockey players based on whether their team scores more goals than the other team,” Dellow says, “and when Grabovski’s on the ice, that happens.”
“One guy, who won a Stanley Cup scoring the game-winning goal in the final minute of Game 6? The other guy’s never been close to that,” Simmons retorts, twice as loudly.
“Yeah, and what did Dave Bolland do?” Dellow teases. “Is he Jonathan Toews’ dad? Because I’m not sure how you’re giving him credit for the team he played on.”
The tension is palpable and very much what we’ve come to expect from a rivalry that exploded over blogs and the social media universe in recent seasons.
“It’s kind of like the Hatfield and McCoy feud,” said Globe and Mail hockey columnist David Shoalts. They went on so long, nobody can remember how or why it got started.”
Little did we know the advanced statistics versus old-guard debate, the nerds versus dinosaurs war, would go from hot fad to revolution over the summer.
It was a full-on NHL takeover for the stat heads. Dellow now works for the Oilers. Sunny Mehta, a pro poker player turned Oilers blogger, was named the New Jersey Devils’ director of analytics. Whiz-kid stat guru-turned Ontario League GM Kyle Dubas, 28, is now the assistant GM of the Toronto Maple Leafs, who also launched an analytics department and poached the minds behind extraskater.com.
The slew of hirings brought validation to the growing community of thinkers who believe possession-driven statistics like Corsi and Fenwick are the best predictors of success in the NHL. And they poured gasoline on the fiery fight emerging between the stat heads and the traditional-thinking journalists.
“They think I’m a moron, to be honest,” Simmons said. “Not someone who’s covered the NHL for 34 years, not someone who coached hockey for 25 and is a level 3 instructor, not someone who ran hockey schools. I’m a moron.”
In one corner, we have the stats movement. Its general position: Metrics like plus-minus are overrated. We should rely less on our eyes and more on possession statistics, which tell us who has the puck more and, in the case of Corsi and Fenwick, which teams or players drive the play toward the other team’s goal more than others. Corsi includes not just shots, but missed shots and blocked shots, as both constitute possessing the puck as an attacker. Fenwick is similar, but without the blocked shots. Tim Barnes, a.k.a. Oiler stats blogger ‘Vic Ferrari,’ points out that advanced statistics help us study the game better, recording specific shifts in puck control and matchups between players better than the eye can.
“The stats aren’t just stats, but events,” Barnes said.
In the other corner: the purists. These are the reporters who believe the best hockey analysis comes from sitting in arenas, sipping hot coffee and observing the actual game. To get the inevitable Moneyball parallel out of the way: the stat crowd represents Billy Beane and Paul DePodesta, and the purist crowd stands in for the old scouts. The latter group champions the tangibility of seeing Bolland score a Cup-winning goal over what Grabovksi’s team subtly accomplishes with him on the ice. Many, like Shoalts, warn us hockey doesn’t lend itself to statistics the same way a sport like baseball does, as the speed of the game and unpredictability make it difficult to track, especially with on-the-fly substitutions.
The old guard is also quick to remind us the stats are young and flawed.
“We’ve seen this before,” Shoalts said. “It’s like any kind of fad that strikes. Everybody goes on the Atkins diet, even though it’s an extremely dubious way to lose weight or get any kind of decent nutrition (laughs).”
In what world, Shoalts asks, can a statistic such as Total Hockey Rating rank Sidney Crosby in the middle of the league? And why, Simmons asks, were the New Jersey Devils bad and the Colorado Avalanche good last season when the possession numbers suggested the reverse?
“Jake Muzzin is the best defenseman in the NHL according to Corsi,” Simmons said. “There are teams he probably couldn’t play for.”
The team Muzzin does play for, however, is the Los Angeles Kings. And guess what? They’ve won two of the last three Stanley Cups, largely on the strength of a dominant possession game. They get exemplary marks in the Corsi and Fenwick department, creating a correlation that’s tough to ignore. That’s why reporters like James Mirtle, who covers the Leafs for the Globe and Mail, have decided no one can turn a blind eye to advanced statistics anymore. While many of the biggest stat proponents are bloggers coming from other industries, Mirtle is among the first high-profile journalists to embrace analytics.
“There’s no reason to be threatened by this kind of stuff,” he said. “It’s not that complicated, it’s not that difficult to understand, it’s not going anywhere. You can incorporate some really simple ideas from it – that possession’s important, ‘here’s the number that reflects possession’ – and it can add a lot of value to the work that you do. It’s not going to threaten personality-driven stuff or opinionated stuff. All it’s going to do is make your work better and help you look at the game a little bit differently.”
Most of us know the nuts and bolts of the debate by now. But why has it become a rivalry? And why is each side so openly venomous toward the other? The first question is what you’d ask your quarrelling children: Who started it?
Actually, the answer is one thing each side agrees upon. It began with the blogging community. Sites like Irreverent Oiler Fans started breaking down and critiquing their teams’ play from a statistical standpoint half a decade ago. On top of analyzing the game, some began critiquing the critics. They’d slam journalists they believed had a myopic view of how the game is played.
“They would take something that someone wrote in the mainstream and tear apart line by line or piece by piece, a lot of times without talking to the person, contacting the person, doing the basic journalistic ethical thing,” Simmons said. “At least in my mind, that’s where it started more than anything else. It’s funny, because if people in the mainstream hadn’t written things, the bloggers wouldn’t have had things to respond to. So one created the other and now you have what you have.”
The advanced stats bloggers had the perfect conditions to gain notoriety and build a rivalry with the old-guard media. First, as Simmons said, his side gave the bloggers material to critique. Secondly, and contrary to what Simmons said, they didn’t have to adhere to traditional media standards.
“Some of the guys on the analytics side, they’re not journalists, right?” Mirtle said. “I don’t think these guys necessarily have to act like journalists. They’re fans.”
Lastly, many of them wrote with a biting sense of humor, most notably the acid-tongued lawyer Dellow, who has shown a remarkably keen ability to get under his opponents’ skin. His own community heralds him as a sort of watchdog who keeps journalists in check.
“No single person has done more to improve the quality of hockey print and broadcast journalism than Tyler Dellow,” Barnes said. “I never think about that when I’m watching a game or reading about it, but I probably should.”
Once the advanced stat proponents developed a strong social media presence, with tens of thousands of Twitter followers, their influence spiked. The sudden prevalence of analytics bloggers-turned NHL team employees reflects that. Simmons insists many GMs in the league are forced to feign interest in possession numbers now to avoid being publicly flogged by the bloggers, and that “guys I know personally who really don’t care much for the stats will say when you ask them on the record, ‘Oh yeah, we do a lot of work in that area.’ ”
With the analytics voice louder than ever, the sniping went into high gear, too. The most high-profile example: a seemingly innocuous Simmons tweet that went viral. One year ago, when the Toronto Maple Leafs started their season 10-4-0, the analytics community predicted the team would fall off a cliff. The reason: the Leafs relied too much on Jonathan Bernier’s stellar goaltending, and their Corsi numbers were terrible. They were getting lucky winning without having the puck much at all. Simmons scoffed at the time: “Good thing the Leafs don’t play in the CHL. The CORSI hockey league. They’re doing just fine in the NHL, though.”
After the Leafs indeed tanked late in the season and L.A., the league’s top-ranked Corsi team, won the Cup. Simmons was a sitting duck. The stat advocates trolled him to kingdom come.
“Do you know many times that’s been retweeted? It’s in the thousands,” Simmons said. “Hardly a day goes by when I don’t get a tweet with somebody chuckling, ‘How’s the Corsi hockey league going?’ That was daily to the point of that moment. Them being 10-4. It wasn’t a broad statement. It’s 140 characters, so you can’t explain yourself.”
Mirtle says he gets bombarded from the other side with just as many attacks, many of them too inappropriate to print, and often picking apart his appearance.
Talk to the stat folks and their most consistent request is for the purists to open their minds and stop fearing new world order. Talk to the old guard and they want the analytics community to lighten up. Simmons and Shoalts say they’re goaded by their rivals, but admit to deliberately doing the same to stir the pot. Shoalts says he simply does it when he’s bored.
“Some of the advanced stat people tend to be really thin-skinned,” Shoalts said. “And you combine that with what seems to be, in broad terms, a sort of condescending attitude. That makes for a pretty explosive combination. That all being said, I do know us and the media tend to be a little thin-skinned, too, which has always struck me as kind of odd, because we dish it out and too many of our members can’t take it sometimes. But I will say this: I find a lot of the advanced stats people kind of annoying.”
Take a bird’s eye view of this war of words and, as is the case with most fights in life, you’ll see the sides relate to each other far more than they realize and even agree on plenty of points without even knowing it.
Look at Simmons’ assertion the stat crowd thinks he’s a moron despite his many years of experience in hockey. Mirtle says his family laughs whenever critics tell him he knows nothing about the game, as all he’s done his whole life is study it obsessively. And the traditionalists paint Dellow as a guy who lives in his parents’ basement surrounded by blinking monitors. Edmonton Sun reporter Derek Van Diest criticized the Oilers’ Dellow hiring, insisting one sees the game differently in an arena, in the press box. The ironic part about that stance: Dellow has spent many a day in a hockey rink.
“I’ve actually played hockey with him a couple of times,” Mirtle said. “He’s a very good hockey player. He’s from a small town in B.C., he’s highly educated, he knows the game very well, he plays regularly. He’s probably a better player than most of the people in the media. We have to get past these petty judgements on people, because whether you like numbers or not has nothing to do with how much you know about hockey.”
Both sides bristle at the notion they’re being called out for not understanding the game, not realizing the very people calling them out suffer from the same frustration. We all know the game, people. That’s allowed.
Even more fascinating: the dweebs and the fossils sometimes argue the same point without realizing it. Check out these “anti” Corsi statements. First, one from Simmons:
“I hear stats guys who have made the case to me that the Chicago Blackhawks win because they’re advocates of analytics. The Chicago Blackhawks win because they have Jonathan Toews, Duncan Keith and Patrick Kane, three of the best players in all of hockey. And why do the L.A. Kings win? Not because they believe in possession. They win because they have the best players. When you have the best players, guess what happens? You have the puck more. That’s pretty modern. You ask Scotty Bowman about the Montreal Canadiens from the late 70s. Why did they win? (a) They had the better players, (b) they had the puck. And there’s nothing better defensively than to have the puck.”
Now, from Shoalts: “Right now, you see these guys on Twitter. I suppose it’s dangerous to paint everybody with the same brush, but there are just a few too many of them for my liking that think they know everything now about hockey, and that we poor slobs who have been around a number of years and still depend on our eyes to tell us what’s good or bad about a certain game or certain player, we’re just too myopic and dim to really understand the brilliance of these stats. Well, OK, from what I’ve seen of the stats, most of them are based on possession. They’re basically telling us whoever has the puck most wins. I’m sorry to disillusion any of the advanced stats crowd, but that’s not exactly new. Syl Apps might have told you the same thing 80 years ago.”
The takeaways: good teams and players have the puck, and it’s nothing new to declare that good teams and players have the puck. And what do advanced statistics attempt to tell us: good teams and players have the puck. Sorry to spoil the fun, but the two sides unknowingly agree. Neither even dismisses the other’s viewpoint entirely. Barnes, a moderate voice who stays out of the mudslinging, says it’s equally wrong to ignore advanced stats and to rely on them exclusively. Shoalts says he’d hire an analytics expert if he ran an NHL team, albeit he literally laughs at the idea of installing one as an assistant GM.
Now that the stat crowd has permeated the NHL, it’s safe to say Corsi and Fenwick aren’t going anywhere. Maybe the old guard and new wave will kiss and make up. Or maybe there’s no onus on anyone to embrace a belief contrary to his or her own.
“I’ve never really understood this powerful need that some guys have to make everyone understand their point of view,” Barnes said. “That’s your edge. If you have a really good edge in business, why on earth would you spend enormous chunks of your time trying to give everyone else that same edge? I don’t get it.”
Or maybe, as one NHL executive suggests, it’s a matter of simple nomenclature. What seems foreign now will be common hockeyspeak in a few years.
“Imagine asking a coach what he considered ‘advanced stats,’ ” the executive said. “Would you say to him, ‘That assistant coach in the press box is an advanced statistician? Because 20 years ago there were none of those guys. If we just called it ‘information,’ would people still be as upset? To try and get as much information as possible?”
Maybe he’s right. Maybe cooler heads will prevail. And if they don’t, hey, it’ll be fun. It sure has been so far.
Matt Larkin is an associate editor at The Hockey News and a regular contributor to the thn.com Post-To-Post blog. For more great profiles, news and views from the world of hockey, subscribe to The Hockey News magazine. Follow Matt Larkin on Twitter at @THNMattLarkin