Tyler Dellow is not the first person to be hired to provide statistical analysis for an NHL team, but his recent decision to join the Edmonton Oilers has set the internet ablaze with hot sports takes on what it means to the future of in-house analysis by NHL clubs.
The Los Angeles Kings have long been known as the front runners in analytics. Their second Cup in three years along with the collapse of the anti-stat Toronto Maple Leafs seems to have finally broken the levee on teams seeking out statistical help. The Leafs hired Kyle Dubas, a young, stat-savvy junior hockey guy as their assistant GM. An unnamed team is working with analytics blogger Eric Tulsky, the Devils hired a poker player to operate their analytics department after an extensive search and now, Dellow.
Dellow’s hire is different though. For years, he ripped the Oilers – with brilliant, in-depth reports and statistical studies and a side dish of snark and venom. He blasted their front office, clobbered coaching staff after coaching staff – and in the end, was right time and time again. They must have finally come to grips with the fact that he wasn’t a PR nightmare, he was a bright young mind who could help them return to being a competitor in the West.
Hiring Dellow, the author of MC79hockey.com (now taken down along with years of great work), came initially with pride. Anyone who writes about stats has been pelted with the “watch the game” and “get off my lawn” garbage that comes along with challenging mainstream ideas. So the Oilers’ decision to hire him felt like we – the stat nerds – won the Cup. We beat the heart-and-hustle crew. Our guy went toe-to-toe with the Old Boys Club and toppled their castle.
But after a few days of reflection, there are some conflicting thoughts on what Dellow’s hire and the flood of teams grabbing up analytics writers really means.
Dellow represented something. The same thing that a bad who says “screw the man” does. He was an underdog who stood for something that only other cool people understood. He gave readers a window into front office and coaching decisions and showed us that those guys really don’t know it all. He was a badass about it too, confronting writers who parrot the same platitudes about winning culture and grit.
It doesn’t quite feel like he sold out or joined the enemy. Yet he certainly is not one of us anymore. And we are worse off.
It is not like the stat community just went into the tank. There are hundreds of bright people doing fantastic analysis. But more of them will be picked up by teams soon and the crew of wannabes will grow larger and larger. It has grown 1,000 fold since I started writing about numbers for WGR two or three years ago. Even the Sabres’ broadcast is showing Corsi numbers after each period. I never even dreamed of that happening.
Still, the hires of top analytics writers opens us up to a drop off in the quality of writing and loss of the guardians of statistical analysis. A writer like Dellow was always there as a filter to tell us whether new studies or ones done by mainstream authors actually made statements about players and the game or if it was contrived nonsense (like, player X hits .300 on day games in June).
We aren’t lost in the woods without him, but it seems the writing is on the wall for things to become harder and harder for fans to discern which statements are logical and which conclusions are off base.
You wonder if teams – who are insanely ignorant about statistical analysis – will hire the right people or snake oil salesman. Dellow is legit. After his hire, you can guarantee 5,000 clowns with calculators sent out resumes. How will your team tell what’s what? The line is thin between a brilliant analyst and a Google Scout.
It was not long ago that I sat across from Tim Murray in a meeting room at First Niagara Center and – along with several other WGR crew members – grilled the Sabres’ general manager about his theories about analytics.
He was right about some things, but others caused an eyebrow raise. He asked the Sabres’ stat gurus to develop a system to rank players 1-7 so he could compare his 1-7 scouting rankings to their reports. It is a nice approach, but there are problems: Brad Boyes was a 3 with the Sabres, a 6 with the Blues, a 5 with the Islanders and a 4 with the Panthers. Numbers in hockey are fluid. Boyes was used differently by each team, had different calibers of linemates, spent different amounts of time on the power play and had different levels of luck.
They aren’t like baseball where a player hits X and it results in Y wins. Every player affects every other player. That's not to say the Sabres are headed down the wrong path, only that: Just because you're using stats doesn't mean you're using them correctly or getting an advantage or that anything someone writes or says that includes stats must be right. Keep that in mind.
A stat person like Dellow dug deeper. He thought critically, challenged ideas, understood scouting tactics and studied strategies. It is not possible to be a good analytics consultant for a team or writer without being able to analyze all elements of the game.
Will your team know how to sort out which is which? Will fans, now that many of the best of the best are being hired away from the blogosphere?
Maybe it is simply an overreaction to a surprising move by the Oilers. Maybe these teams will hire bloggers to shut them up and stick them in closets instead of utilizing their knowledge. They have to protect their own, after all. Crazier things have happened. Major League Baseball teams hiring guys from Baseball Prospectus has diluted the product, but it hasn't ruined baseball analysis.
It hasn't rid us of bad decision making by front offices or of the Oakland A's being ahead of the curve. Using stats may lead to a few more career AHL'ers landing in the NHL to replace checkers and maybe a strategical change similar to shifts in baseball. Contracts and signings will be changed but fans probably won't see a notable difference.
Now that stat guys are in the front office, we may never see the next wave of innovation such as player tracking. That's too bad. The NBA has all their tracking stats at NBA.com but I'm not confident the NHL will allow it out as MLB did with PitchFX and the NBA has done. We'll see. I don't expect a paradigm shift just yet.
Anyway, good luck to Dellow and the others. If we get nothing else out of this, the hires seem to be driving some mainstream writers nuts - and that’s pretty cool.