Narratives in the analytics era

The analytics movement is gaining steam, but the traditional narratives still appeal to a broad audience. What changes are we seeing in mainstream writing?

By Eric T. Feb 4, 2014, 2:46pm EST

If you buy something from an SB Nation link, Vox Media may earn a commission. See our ethics statement.

Artist's rendering of a blogger and mainstream writer discussing hockey.

Non-traditional statistics are starting to have an impact on the mainstream media.

National writers like Elliotte Friedman are weaving Corsi data into their columns and televised broadcasts are giving Corsi updates. The NHL hired Olivier Bouchard, Sportsnet hired Tyler Dellow, Neil Greenberg writes for the Washington Post, and the Globe and Mail has given James Mirtle the freedom to really dig into what the stats say about the Maple Leafs.

But a lot of people aren't interested in statistical analysis, and the mainstream outlets need to appeal to a broad range of people. So while I expect to see this growth continue, it will have its limits; there will always be a place for well-crafted narratives.

I've been thinking lately about how the changing landscape of sports writing -- and especially the rise of analytics -- might affect those pieces and their writers.

The hot start

Colorado started the season 12-1.

Last year, their 39 points and minus-36 goal differential were second-worst in the NHL. They haven't won more games than they lost or made the playoffs since the 2009-10 season. But they brought in Patrick Roy to be their coach and then started this year 12-1.

If you're a writer who focuses on the human element of hockey, this seems like a story just begging for your attention. The legendary goalie returning home, the young kids being inspired and committed, and the goalies posting a .955 save percentage that would make even Roy proud.

Two years ago, the Minnesota Wild changed coaches and had a similarly hot start. Both team bloggers and national writers embraced the idea that the results were the product of the team embracing the new coach's system.

Yet I saw very few go down that road with Colorado. Obviously, people noted their hot start, but I saw very few people making the jump to concluding that they were a dramatically improved team.

So what was different this time?

The challenge for writers

Let's be fair: this was just one story, and there doesn't have to be any specific reason it didn't get written up more -- it could be that other stories just happened to catch people's attention instead. But whether there's a direct causal linkage or not, the lack of excitement around the Avalanche's hot start got me thinking about how things might be changing.

Coming up with explanations for these short spurts is a staple of the traditional view of sports. These explanations -- things like chemistry, momentum, system play, confidence -- almost always derive from the human element of the game in some way. This is how most people think about sports; these are the stories most readers want to read.

So a publication serving a broad audience will almost have to include some of this work.

The statistically oriented tend not to go down that road much; the reluctance to read a lot into relatively short stretches of play is one of the biggest things that distinguishes statguys. As the analytic approach spreads, so too does the vocal, perhaps obnoxiously self-assured group using blogs and social media to dispute and debate these traditional perspectives.

As the evidence mounts that there's truth in what those stats tell us, that the Wild or Avalanche having an extraordinary save percentage indicates a transient run of good luck rather than a fantastic new system, writing engaging but accurate articles for a broad audience becomes more difficult.

I reached out to a handful of writers for national and local mainstream outlets to see how they are walking this line.

Public scrutiny

One thing I wondered is whether there really has been a change. The circle of people I interact with certainly isn't a representative sample, and maybe I had a skewed perspective of how much impact new media are having.

Anthony SanFilippo has covered the Flyers for the Delaware County Times and the Flyers team website. He considers himself a pioneer when it comes to the use of social media in sportswriting; he says that in 2009, he anticipated what was coming and became the first Philadelphia sportswriter to use Twitter.

He believes that the internet has had a big impact on sportswriting.

"We were all trained to get good stories and we never had to worry about being fast with it because it wasn't being released until the next day. Then when you had to start putting stuff out right away using social media or a blog or whatever, getting it online, it became a rush to judgment.

"Just get it out there as quick as we can, and you know, we're going to be wrong sometimes. But that's OK, we can always correct it online."

He notes that this led to scrutiny, to people watching and catching those mistakes, and that this has been hard for some writers. Newspapers weren't in a position to turn things over to social media experts as things changed; they just expected their existing, well-established writers to adapt, and that hasn't been easy for everyone.

"The narrative has to rest on the facts."

Nick Cotsonika of Yahoo! Sports considers himself a writer focused on the human element. With his goal of appealing to a broad audience, he is keenly aware that using a lot of stats can weigh down the writing. He didn't say as much, but I didn't get the impression stats will be a focal point in his writing any time soon.

Nevertheless, he finds that "over the last two or three years especially, statistics has made me rethink some of what I'm writing. It's made me use a little more discretion in what I write."

He says that writers are wired to look for stories, and that unusual streaks often stand out as newsworthy. While people joke about jinxes, he realizes that it's often just simple regression.

So while he's looking for those interesting stories, he's keeping in mind that "the narrative has to rest on the facts. Even if you don't include a lot of statistics, it has to be solid information."

Taking cues from teams

Elliotte Friedman occasionally includes a direct reference to a non-traditional stat in his column.

He considers being up-to-date on what's out there an important part of reporting. He's not yet sure exactly where the truth lies with respect to stats, but he's trying to learn everything he can.

"If you want to be the best at your job, you have to try to understand everything about it."

He's heavily driven by what the teams themselves do. When he covered the Raptors, they told him they considered rebounds per minute to be an important stat, so he started looking at it. Regardless of whether it's actually a good predictor, if the team valued it, he could use it to help understand and interpret their actions.

He says it's hard for hockey stats to get really widespread acceptance because no teams are willing to discuss what stats they use. Several of them are known to be heavily involved in stats, but they are very secretive about their work.

Presumably, the lack of a clear signal from the teams leaves beat writers without a strong incentive to investigate or discuss stats in their analysis.

Analytics in print

No team has given a more clear signal about their position on statistics than the Maple Leafs.

While he was GM of the Leafs, Brian Burke openly scoffed at the value of statistics. Greg Cronin scoffed at the idea that shot differential might relate to puck possession last off-season. Dave Nonis says that the Leafs set aside money for analytical work every year and so far haven't found anything worth spending it on.

And yet as a beat writer covering the Leafs, James Mirtle has taken a very analytical approach to understanding the Leafs.

The team may not have led him down that road, but he says he's always been an analytical person. He acknowledges that some of what the stats show isn't intuitive, but says that you find their real value as you see things play out in the NHL over a number of seasons.

Sharing that with a broad audience might not be easy in every situation, but Mirtle is in a great situation to make it work.

His newspaper, the Globe and Mail, is heavily focused on business and so has a relatively analytical core audience. The paper has been very supportive of his approach, from the editor-in-chief down.

The response from readership is generally very positive; he senses "a real hunger for less cliched, narrative driven coverage" and notes that the traffic numbers back that up, that while there's always going to be resistance from some corners to something new, there is a large audience that really enjoys getting a different perspective.

My take

The hockey journalism landscape is changing for a variety of reasons.

As analytical work continues to catch more people's attention, I hope and expect to see more mainstream writers taking Mirtle's approach, appealing to the growing audience interested in what the numbers say. We're already starting to see a couple of publications add this kind of work to their repertoire, and I think that trend will only continue.

But there will always be a strong market for the sort of human element story that Cotsonika writes; those pieces should and will continue. All I would ask of those writers is that they take Cotsonika's words to heart, and try to make sure their work is grounded in fact.

For me, the dividing line is between past and present.

That hot streak really did happen -- the shooter really did pick the corners on every shot, the goalie really did make some amazing saves, the team really did play the coach's system perfectly. There's nothing wrong with writing about that.

But nobody's shots stay wired forever, and whether the goalie was locked in yesterday doesn't appear to have much bearing on how he'll do tomorrow. If you find yourself crossing that line from description to projection, you're getting into an area where the analytical approach generally does better than the emotional angle.