The book that first made me want to write about soccer—as opposed to just watching the living hell out of it every weekend—is currently sitting on my desk as type this sentence. It’s dog-eared and yellowed, and the cover is pretty frayed. It’s also dated; the copyright date is 1990.
When my brother first gave me Brian Glanville’s The Story of the World Cup many Christmases ago, I didn’t think much of it; just another soccer book, a last minute, used bookstore gift-grab. That of course all changed when I finally sat down to read it. Here, for example, is Glanville’s take on England’s 1966 World Cup win:
Once more the Germans threw their men into unrbridled attack; once more their defence broke down in consequence. It was in the last seconds that Moore’s long pass, capping an immaculate performance, sent Hurst through, and this time there would be no erring. As joyful small boys dashed on to the pitch, anticipating the goal, the final whistle, he carried on alone, blew out his cheeks, and beat Tilkowski with a terrible left-footer. He was the first man to score three in a World Cup Final; and the Cup itself had at last come home to the country where football began.
As a Toronto kid brought up on the terse, one-sentence paragraph of the newspaper beat writer, this was inspirational cocaine. I mean, you simply weren’t allowed to write about sports this way, what with romance, story-telling, art, were you? Who wouldn’t want to write about football, or indeed all sports, like Glanville? Why weren’t more sports writers in North America doing this?
That sense of romance hasn’t eluded me, but something has changed in the way I understand the football journalism in the last year or so. I don’t know exactly when my interest in sports statistics began, but I’m pretty sure it involved writing in close proximity to the sabermetrics-friendly folks with Getting Blanked when I first joined the Score over a year ago.
What first interested me in sports analytics—far, far more than the prospect of gambling-friendly predictive metrics or the lure of transfer market-defying ‘Moneyball’ methods—was its emphasis on empiricism. X team beats Y team: why? What exactly happened that led to the outcome (other than goals of course)? What can teams do to win games in a way that breaks away from regression to the mean?
I’m not a statistics person, so my interest is still largely from the outside looking in. But I do understand, at its core, the beautiful simplicity of statistical power. And in some of the more egregious cases, I can tell the difference between ‘good’ and ‘bad’ analytics at this early phase. But that interest had a significant effect on how I read football punditry on blogs, newspapers and media sites.
While I’d always disliked simplistic explanatory cliches in a lot of soccer writing, analytics put them into sharp relief. Did teams really lose title races because of a “lack of mettle”? Was the English national team’s failure really down to “too many foreigners in the Premier League”? Did a defensive midfielder really fail because he “didn’t tackle enough”?
I used to think the suspicion among mainstream football writers with regard to analytics had to do with a fear that a game that falls within the bounds of empirically and statistically accurate probabilities is a game without the key element that makes all sports compelling: uncertainty.
Today, I’m more inclined to believe the suspicion of analytics involves the fear that the entire edifice of football punditry—which right now resembles metaphysical castle in the sky built on cliche, personal and national bias—would be held to a new, frightening standard: empirical fact. That would mean writers who once comfortably assumed a sense of authority based on their media platform might be held accountable for what they say.
If that sounds impossibly far-fetched, consider yesterday’s result in the US presidential election. Last night I was far less interested in the result for its political implications than I was for the reputation of fivethirtyeight’s Nate Silver.
Silver, as I’ve written before, is an interesting bellwether for the sports analytics movement (he got his start in sabermetrics with PECOTA). His blog, with regularly updated election probability forecasts, didn’t use a complex, hidden formula (he’s not a statistics professor), but rather aggregated all major state and national polls, corrected for historical bias and other extraneous factors (incumbency, the economy) to produce a simple win/loss percentage (along with a host of other probable events).
Throughout the election, Republican challenger Mitt Romney never once led Barack Obama in Silver’s prediction. On the eve of the final vote tally, Silver gave odds to Obama in every battleground state, with Florida a solid tossup. Obama’s reelection odds held at 90.9%. As Silver repeatedly stated (to apparently deaf ears) that the only way Romney could win was if the individual polls were all statistically biased against Mitt Romney, perhaps involving a massive misreading of the 2012 electorate based on 2008 numbers. He also showed with rigourous accuracy that the poll numbers did not in fact signal any momentum for Mitt Romney in the final stretch.
The key point to take away from this is that for Silver, the election was never a tossup. Silver’s accuracy in previous elections, his solid empirical methods, and his emphasis on statistical accuracy made him a figure of derision by pundits on the right of the spectrum, and a subject of weird, almost guru like curiosity on the left. Many came up with their own electoral vote counts based on what were essentially “hunches” based on demographics, expected turnout, Hurricane Sandy, etc. And while some pundits took egregious issue, most if not all major news outlets referred the race as a “dead heat” when it was demonstrably nothing but. Slim perhaps, not assuredly not a tossup.
So who was right? Take it away, Business Week:
Nate Silver was right. The Gallup Poll was wrong. Silver, the computer expert who gave Obama a 90 percent chance of winning re-election, predicted on his blog, FiveThirtyEight (for the number of seats in the Electoral College), that the president would receive 51 percent of the popular vote as he called each of the 50 states, including all nine battlegrounds.
As far as statistics are concerned, Silver isn’t some sort of wiz (the worst thing would be for him to take on a Neil de Grasse Tyson-like status among admirers of his methods). He simply applied basic statistical poll averaging with some historical provisos to produce a number. He did this on the clear evidence that polls averages were broadly historically accurate.
Yet it wouldn’t be a stretch to say his approach in some sense made the entire edifice of predictive electoral punditry redundant.
We’re currently nowhere near a similarly accurate predictive model in football analytics. But we already know that some of the statistically simplest variables to isolate over time are also the most effective in determining which team is likely to win things in the long term. Statistics need not be some shadowy science.
Note that none of this means the romance of good prose writing about football is dead. Uncertainty remains intact. All the things that make football great won’t disappear with advancements in analytics.
But it should be a sign of the beginning of the end of lazy, factually inaccurate football punditry posing as authoritative prose from so-called, self-appointed careerists selling their ‘expertise.’
That end by the way won’t come as a matter of moral principle, but inevitability. As I walked home last night, I listened to NPR call the election results. They simply listed certain states were too close to call, and repeated the same senate wins over and over again in between various segments on what the “mood” was like in various campaign headquarters. They did what mainstream news networks have been doing for years.
I got frustrated and instead visited the fivethirtyeight live blog, where I learned which county in which specific battleground state had reported lower numbers for Romney than McCain in 2008, and that these county’s were traditional bellwethers for those states. I read Nate Cohn’s blog to see the shifts in demographics based on tenuous exit polling.
In other words, I skipped the fluff and went to the juice. There is precious little juice in football analytics at the moment, but the day might come when football readers will tire of reading the same old tired cliches in various papers, and visit a great analytics website to determine what actually happened.
I believe finding this “juice” is the future of sports journalism, and I’m proud to work with a company whose group bloggers long ago recognized the value of empiricism over blind faith, and know the difference between gorgeous prose and meaningless cliche. There will always be Brian Glanvilles in football writing (I’m looking at you Brian Phillips!), great writers who can capture the ineffable, poetic glory of the game. Rather, the best writers will recognize that the real romance of football is hindered by factually inaccurate and lazy cliches, not helped by them.