Dear StatsBomb: Part 1
The European season is over. The World Cup is still two weeks away. We have entered the Great Content Gap. But never fear. I shall step in to the breach and fill the yawning hole in your football media consumption diet with my first ever StatsBomb Mailbag.
Let’s kick things off with a question from the boss.
Where do you want the content side of StatsBomb to be one year from now?
Because Ted’s the boss his question goes first, but because I run the website I get to refuse to answer his other question about what my most painful moment as an Everton supporter was.
What’s interesting about trying to answer this, other than it being essentially an interview question that I’m now answering in public, is that StatsBomb has a lot of freedom in terms of direction. As I’ve written about before, the website exists because the greater StatsBomb data collection and consulting company exists. We are lucky to get to build on the data and not be beholden to selling ads to justify our content.
So, if the purpose isn’t to chase clicks (and it isn’t) then how do you measure growth, and channel ambition. Well, one axis is simply the quality of the work. A year from now I want Statsbomb to be a place where readers know they’ll be getting great content, and writers know they have a home for smart ideas. Just because we aren’t selling ads doesn’t mean we’re indifferent to getting eyeballs. We want people to read StatsBomb, it’s just we want them to do it because they believe in our product and our content, not because we have a business model that depends on it.
That doesn’t mean that great content can’t also make money. It wouldn’t make sense for StatsBomb to become a subscription site (since the entire point is to be a free place where people can read good work which is supported by the resources that the services StatsBomb actually do sell) but that doesn’t mean there aren’t other ways to grow a business. One particular idea that has always intrigued me is trying to make and sell a Baseball Prospectus style season preview, perhaps with periodic email newsletter updates over the course of the season for those who purchased it. It’s a good engine for developing a discrete marketable content product but also not cannibalizing the free content we’re committed to giving y’all.
A year from now I hope that the content on the website continues to grow, evolve and improve (and maybe increases in volume) and that on top of that we are launching different types of content products, some of which are even intended to be profitable.
Oh, and I also want a copy editor.
This was such a good question that Ted decided to answer it on Twitter before I could even write the mailbag.
The question and answer here get at something I think is often misunderstood about analytics and the nerds that do it. Frequently analysts will be skeptical of causal claims surrounding soft factors. Things like confidence, or succumbing to pressure in big moments, or any of the other clichés that swirl around the idea of mentality. But, crucially, that doesn’t mean that they don’t exist.
Human beings are human. Sometimes they feel good and sometimes they feel bad. Sometimes they perform well. Sometimes they perform poorly. Sometimes my articles contain lots of typos and sometimes I write clean copy. For data driven people the skepticism isn’t that the underlying thing can't be true, it’s the assumption of causality. Analytics is a field that’s deeply aware of the idea of uncertainty and randomness. It’s not that a player can’t be clutch, it’s that a player who isn’t clutch in a big moment would look more or less the same to an observer as one who had a four-month-old baby that kept him up half the night, or one who was fighting off some bad pad thai or one who just plain old got unlucky. Claiming to know that it was specifically a weak mentality (or whatever you want to call it), as opposed to a weaker claim of simply acknowledging that as one of a number of possibilities is where the difficulties lie.
Analysts are rightly quite reticent to make those links strongly given the relatively little data we have. And, quite frankly, historically we aren’t great at correctly predicting who is bad in a big moment and who isn’t. Players choke until they don’t. Arjen Robben was a failure in big moments until he scuffed a shot that found its way into the back of the net in a Champions League final. Then, he isn’t. It’s not that soft factors don’t exist, it’s that all too commonly the things that end up defining a player’s mentality are the things they don’t have control over as opposed to the things they do.
It’s an important distinction. It defines how analysts look at a problem. Ted’s response is typical of good work. Be skeptical, but look for data. Try to ground the problem as much as possible. Accept both the reality that humans are humans, and that we are terrible at understanding the transmission mechanism from the human brain to performance on the field. Then, work to bridge that gap.
Ok, now it’s time to fill this particular precondition for doing a mailbag.
Listen, the obvious answer to this question is Jose Mourinho. The man is a born troll. He’s a master at deploying the same arguments against adversaries that he ridicules as absurd when deployed against him. He says things that are technically correct while being wildly inaccurate if contextualized. He is brilliant at taking any argument and twisting it around until its fought on whatever grounds he wants to fight it on. When he’s feeling threatened he bursts into the football world’s mentions and refuses to leave.
Of course he’d have burner accounts. Mourinho would have a gazillion sock puppets. He’s already the king of turning any fan base that rallies around him into an army of trolls repeating at ever ascending volume whatever dubious claim he’s making about why this week’s disappointing result wasn’t his fault. Jose Mourinho is the king of the trolls, burner accounts are par for that course.
Injuries and analytics are a combustible brew. There is definitely both a desire for, and a realistic possibility of, advancements through data which can help both spot injuries and prevent them. There’s also a lot of quackery. And, because we’re dealing with science and medicine and data, it’s even harder of non-experts to determine the difference between the two.
On a general top down lever there are things teams can do to mitigate injury risks, and lots of the sports science industry is dedicated to developing injury prevention methods. Whether that’s monitoring work load (and having Arsene Wenger ignore that monitoring and play players who are in the “red zone” anyway) or identifying when movement mysteriously drops off, or advising on sleep cycles and eating habits, teams across sports certainly try and ring every last drop of an edge that they can get out of the medical side of data.
How much any of this works is an open question though. And in a specific case, like Nabil Fekir’s surgically repaired knee, it’s even more questionable. There’s no way that from the outside anybody is going to know more about Fekir’s health than a good sports medicine team working with him personally. The strongest claim that you might be able to make is that data could eventually be used to raise the kind of red flags that might make you leery enough to want to get a second, third or fourth opinion.
I founded my own short-lived blog in April of 2013, and started writing for Grantland in June of that year. I think what’s surprised me the most has been the distribution of work done in public vs private since then. In American sports a lot of work was done publicly before teams began to take it seriously. Baseball work was around for a long time, and it took a more widespread adoption by the public before teams began committing to using that work to gain an edge for themselves.
In football teams, or companies that make their money consulting with teams, stepped into the analytics world much more quickly. The pipeline from talented blogger to super-secret team consultant happened faster than it did in American sports. It actually happened faster than public awareness of analytics happened. Despite all the usual silly battles over real football men, and air conditioned offices, owners have been fairly quick to hire analytics people.
Now, that’s different than saying that whoever they hire as any real influence, or that team transfer policy, or management choices have reflected analytics best practices. Managing teams are tricky, effecting change is difficult, and doing analytics well is hard. Put all that together and you’ve got an environment where hiring people is the beginning of the process not the end of it. But, in general I’ve been surprised that the speed of teams doing at least some analytics works seems to have outstripped the speed with which work is viewed as mainstream by fans.
Have I been a big enough influence on your writing that you’re going to stretch this mailbag content into a two day thing?
--Fake Bill Simmons
You better believe it.