NHL: Is Hockey Analytics Growing Too Fast?

The trend of advanced analytics seems to be taking the NHL by storm, but can it quantify everything?

Last week I was sitting in my evening economics class when our professor asked us an interesting question. He proposed, “if baseball players had the freedom to sign wherever they pleased, do you think all of them would play for major markets such as New York or Los Angeles?” This can be replaced with any sport, including the NHL.

The topic centered around the concept of marginal cost and marginal benefit, or the principle that measures the value of additional units (or players) weighed against the cost of signing them. My professor also brought up if this idea would be good for the league.

I raised my hand and brought up a bunch of stupid baseball knowledge like parity, revenue-sharing and entry level contracts. He swiftly refuted saying none of those things mattered in this situation.

After thinking about it for a moment, I said they all would sign in major markets.

I was wrong.

From an ownership’s point of view, the cost of adding additional superstars to big, long-term contracts would eventually catch up to the benefit of having another great player on the team. Therefore, the competitive market would eventually spread the wealth of talent around the league.

That night, I got home and started to think about hockey. I thought about the Washington Capitals trading for Kevin Shattenkirk and the Nashville Predators trading Shea Weber for P.K. Subban. Then I thought about Zach Parise and Ryan Suter signing their monster deals a few years ago with the Wild. I tried thinking about the ways free agents and traded players impacted teams and if those marginal benefits outweighed the marginal costs.

It’s a hard concept to grasp at a personal level and an even harder concept when managing a professional hockey organization. But if I learned anything from business school, it’s that every action in your organization must be moving towards the overall goal to be productive. Which finally leads me to something I’ve been thinking about for a while – advanced statistics.

More from Puck Prose

I’ve been fascinated with advanced stats for a little while, trying to read Rob Vollman’s books when I can and checking out the occasional hockeygraphs.com page. It’s always baffled me that things like Corsi don’t seem to really interest old school hockey guys.

I recently listened to a podcast by Theo Epstein back in January where he discussed some of the tests he and the Cubs put their prospects through before the draft. The Cubs teamed up with a team of neurologists and designed a software that tested reaction time. Epstein noted that Kyle Schwarber scored very high on this software, which influenced their choice in drafting him a lot higher than he was projected.

In sports like baseball, where production can be focused individually, advanced statistics can be better applied. However, in a sport like hockey where five players are constantly moving, it’s harder to take those statistics and assume they would be the same in another environment.

For example, say Nolan Arenado from the Rockies gets traded tomorrow to the Marlins, his hitting and fielding abilities would probably be similar. You’d likely see a drop in production, but it would be beyond his control. Conversely, if Seth Jones gets traded from the Columbus Blue Jackets to the San Jose Sharks, his advanced statistics aren’t necessarily going to translate.

Jones is a great defenseman and plays a lot with Zach Werenski. He also plays in front of the best goalie in the league. His statistics are also heavily influenced by the system in Columbus. Jones has experienced the feeling of several variables changing around him. He was still able to adapt and continue to be successful in Columbus.

But there are always young players with high expectations that crash and burn mysteriously. Success in one environment never guarantees success in another. Which is why having a plan in place and never straying from that plan is essential to winning in sports.

I did a little research a while back because I was curious as to what it takes to win the Stanley Cup in the salary cap era. From 2006 until today, I measured things like winning percentage, if the coach had Stanley Cup experience, if they won any of the regular season awards, etc. Everything looked normal with the exception of three categories that stuck out.

Of all the teams that have won Stanley Cups since 2006, 18 percent of them had the Norris Trophy winner, 18 percent of them had the Selke winner and 27 percent of them had the winner of the Mark Messier Award.

In his article (and later book) The Seven Leadership Secrets of Great Team Captains, Sam Walker of the Wall Street Journal talks about the intangibles of some of the greatest leaders on the best teams ever. Some obvious, some not so obvious.

But how can you measure leadership? How do you know that guys like Jonathan Toews or Sidney Crosby had a great night even if they scored no points? How can you justify all the “little things” they do for their teams?

The biggest obstacle the NHL is dealing with regarding advanced statistics is, ironically, what might not be possible to quantify – intangibles. Because hockey is such a team oriented game where intangibles are so important, the least selfish players are sometimes among the most valued.

So all in all, do I think analytics will work in hockey and the NHL? Sure, numbers are a great place to start building your case to make an educated decision. But is winning as easy as a math problem? Absolutely not. There is no magic formula to winning Stanley Cups.

Next: Top 25 Centers In The NHL

As the years go by, the NHL will most likely shift back to a size trend and speed will be a thing of the past because strategy is cyclical. NHL teams tend to follow the leader until originality creates a new trend. But until then, my plan is the embrace the hopefully revolutionary trend of advanced statistics and hope it has a positive impact on the game we all love.