Michael Moynihan, Web Summit 2014

David Epstein got the Sports Stage section of the 2014 Web Summit in Dublin off to a terrific start yesterday with a strong plea for the usefulness of small data in sport.

At a time when there’s a growing fondness for huge volumes of information and statistics – big data – sports scientist Epstein made a passionate plea for focusing on little details, and gave good examples of how doing so can make a huge difference in sporting performance.

Offering the shot putt as an example, he instanced three determinants of performance as release velocity, height of release and angle of release, and pointed to a competitive example where the only difference between the top two performances was one degree in the angle of release: by releasing the shot one degree higher one competitor secured the gold medal.

“I think that struggle, between big data and little data, is playing out in sport generally, which is not to denigrate big data at all,” said Epstein. “If you go back to the ‘Moneyball’ era there was a lot of low-hanging fruit there to be gathered, but the problem was that while sometimes coaches resisted, there were other occasions when coaches jumped right into that without bothering with the basic science.

“If our interest is in helping people improve in small margins, I don’t think all of that can come out of big data. People have said that math is taking over science because you can take the patterns out, which I understand — you can do things with big data you haven’t done before — but it’s a problem that has caused some skipping over – skipping over basic physiology, skipping over skills acquisition and basic knowledge about learning.

“There’s a lot of that which still has to be learned – defining what matters, and what has to be changed, but there’s a well-intentioned fervour for big data which can eclipse some of the other stuff.”

Epstein also questioned the much-touted 10,000 hour rule, which presupposes that expertise in sport (or music) is predicated on 10,000 hours of practice.

“That’s been linked to damaging early hyper-specialisation in sport,” he said. “Instead of kids sports-sampling – take Roger Federer, who played a lot of different sports before settling on tennis – they’re focusing too early on in life on just one sport instead of sports-sampling.”

As a consequence, he pointed out that statistics prove such hyper-specialisation has drastically reduced children’s chances of succeeding in making it as professional sportspeople.

The author of The Sports Gene added that a recurring issue in modern sports is a basic lack of expertise in statistics. Accepting the 10,000-hour rule, for instance, can be a result of focusing on just one example of a sports prodigy: “One huge problem, not getting too wonky about it, is the restriction of range problem. A journalist – or scientist – will start, say, with only elite performers, which restricts the range and can cause all sorts of problems.

Epstein also pointed out that writing about genetics and race can be a fraught business, with so many observers swift to take offence at what they perceive as racism.

“Writing about race and gender and so on, you can be nervous about what may happen, and writing my book I spoke to scientists who told me they were hiding data on genetics because they didn’t want to be involved in controversy.

“To me the way to proceed is to find out what differences between people are real based on science, not on intuition, and to find out which of those difference actually matter, and how you can use those to find the best outcomes for all people.

“We’re fine using medical genetics that way, but once we get outside that area, we’re ‘oh no’, which to me is a derogation of duty.”