Search

My Mind Has a CF% of 35%

I've been asking myself the same question repeatedly over the last year or so, and I'm struggling to answer it. Why do I despise "Fancy Stats," otherwise known as advanced statistics and their use in hockey?

For those who are unaware, advanced statistics go deeper than goals, assists, points, plus-minus, shots, penalty minutes, etc.

Examples like Corsi-for percentage (CF%), expected goals percentage (xG%), expected save percentage (xSv%) and many others help calculate things from puck possession, shot quality, goaltending performances, defensive performances and even how lucky a player or team might be.



My face when I try to wrap my head around xG

"Ovechkin Stretches"by clydeorama



Those things are important. They were refined by people who formulated hypothesis, experimented with them and refined them through discourse to be as accurate as possible. Some would call it hockey's scientific method.

I am not anti-science. I wear a mask when I leave my house to avoid spreading or contracting COVID-19, and I know global warming is real. I'm not a math whiz but I know its importance, so why do I hate advanced statistics in hockey?

Why do I start getting that weird twitch when I see a tweet or an article referencing xG% or CF% or any of the others?


Before you close this tab, here is an explainer for many of the advanced stats I will mention. I won't blame you if your head starts to feel like you just ate an entire ice cream cone in seconds. It can be headache-inducing.

So, why do fancy stats bug me? It's not that they are used in hockey, but how the people that cover the sport use them.

Jocks versus Nerds

I'm not one who enjoys confrontation, on social media or otherwise. When I do find myself doing it or fighting the urge, it's almost always due to someone's use of an advanced statistic.

The immediate response I almost always try to stifle, because it's not right, is the "I played hockey, so I understand it, you didn't so you don't," response.

I did play, but I didn't play in college, and I didn't play professionally. So, to me that response is like Uncle Rico telling you how far he could throw a football in high school.

There is no hard and fast rule that says if you didn't play that your analysis or your point of view isn't important. Anytime a pro athlete says that to a reporter, they get killed for it and rightly so.

But why do I have that reaction?

Advanced stats are important in hockey. I believe they can help a general manager build a team (although they shouldn't be the main focus). I believe they can help scouts and coaches concretely measure a young player's improvement if that player struggles to put up points for a time.

There are many advanced statistics that I understand and just as many that I don't.

That's the problem.

The disconnect comes from within my profession, sports journalism. One of the golden rules of sportswriting is to "explain things like you're explaining it to your grandmother."


The idea is that you want to appeal to a broader audience; not everyone who reads your writing knows hockey, or math as well as you do.

Many of us who write about hockey now have thrown our grandparents under the bus.

Ethel and Ernest Ciardelli had the fantastic privilege of driving from Southern New Hampshire to Boston to watch Bobby Orr become a legend. But they would not have understood CF%, xG%, or let's be realistic, anything that was invented after the year 2000.

But you get the point; many hockey writers aren't using fancy stats correctly in that they fail to explain or provide any context. You could argue that the people who seek out hockey writing outlets have a strong base of knowledge, but I think you'd be shocked.



My grandparents, Ernest and Ethel Ciardelli used to watch this guy play in person. They understood hockey. They most definitely would've need an advanced stats explainer if they'd have been able to figure out a cell phone first. (Ross Dunn)



Please don't go to Twitter and show me how many followers you have or how many retweets or comments you get.

How many people engage you in the part of your tweet that references an advanced statistic? Could it be that they’re afraid to look stupid because they don't understand?


If you forced the majority of hockey fans to explain how CF% or xG% or any of the other advanced statistics are calculated, I bet they couldn't tell you, which is essential.

When you are referencing them in an article or a tweet without context or explanation, you are doing them and yourselves a disservice.

Natural Stat Triggered

The event that motivated this blog occurred last week.

A fellow hockey blogger, who I like, posted a tweet that referenced how well a player currently playing in the NHL playoffs had performed in that game, citing only his xG%.

Here's the problem; that player had a miserable game. He performed so poorly that local beat writers called for his benching. It wasn't just an insider's opinion, he was objectively terrible, yet anyone who had read that tweet and hadn't seen the game would've had the wrong idea.

It made me wonder, did this blogger even watch the game themselves? Or did they visit naturalstattrick.com afterward because they didn't have time to watch the game and the player in question didn't show up in many highlights? That's what I think happened.

But, if that writer truly believes that the player in question had a good game, they should explain themselves beyond posting an advanced statistic!

What about that player's performance resulted in that statistic? What did he do well?

In the case of expected goals percentage, was that player the beneficiary of rebound opportunities off shots from linemates who happened to take a lot of shots? Is one of his linemates an elite passer? Is the player himself good at finding unguarded places in front of his opponent's goal where he can get to more rebound attempts? iIs he lightning fast on the rush making him more difficult to defend and more likely to generate quality shots?

Also, does the specific advanced statistic take every variable in a player's performance into consideration? Odds are it probably doesn't.


Advanced statistics exist to break down the game into small parts and measure things that might not be obvious. No advanced stat can definitively tell you that player x had a good game without being put into context.

In the situation I'm thinking of, the player made two errors in his zone that resulted in two goals against his team.

His was, indeed, benched the next game.

Anyone who reads that tweet without knowing the full story themselves now has the wrong idea about what actually happened. It's not life or death, but as a hockey writer, you didn't do your job, and it's a mistake many of us make. You are relying on an advanced statistic when there is far more to the story.

It brings the "eye test" and the "math test" together because those two schools of thought shouldn't be on opposing teams in a battle to see which is more valid, they should be linemates helping to achieve the goal of explaining the game of hockey to many more people.

12 views