Marginal gains’ growing pains
What particularly strikes me is that all of his examples of marginal gains in action relate to machines. Cars at Mercedes, computers at Google, bikes at Team Sky. He mentions in passing that the ‘data-driven’ approach to improving the performance of these machines through ‘testing and adaptation’ can be transferred to other aspects of elite sporting performance, such as ‘diet, training methods…recovery protocols and many others’. But what he fails to acknowledge is that these ‘dimensions’ of sport are entirely different from, say, a bicycle’s ‘aerodynamic efficiency’. In short, rather than relating to systems designed and created by humans, they relate to humans themselves. And at no point is any evidence given for the benefits of marginal gains when trying to optimise the performance of a biological system.
In fact the only concrete example given of a tactic to help ‘get ahead of the game’ in physiological terms is doping. Syed mentions it in the context of Team Sky’s success, of course, to dismiss it (with minor caveats). But it’s striking that he gives no other example of a ‘marginal gain’ that would actually improve the physical performance of a human being. Nevertheless, he insists that even if Sky did engage in doping, the beautiful marginal gains baby need not be thrown out with the dirtied bath water. Hilariously, he even equates doing so with disregarding experimental science(!) on account of a few data fabrications. One can’t help but feel that this sort of overstatement might be what happens when you have talks to give and books to sell.
Syed can assert the life-changing significance of marginal gains as much as he wants, but the question of what, exactly, is new about it remains. In engineering, which is what he spends most of the article talking about, I would posit that the process of monitoring the performance of different parts, and adopting a policy of ‘testing and adaptation’, is not exactly revolutionary. I took an introductory programming course last year which emphasised, repeatedly, the importance of putting a structure in place for efficiently testing your code, of then testing it after nigh-on every change made, of making your code modular so errors in individual parts are easier to correct and don’t impact too heavily on the rest of the system (which sounds unnervingly like the concept of ‘isolating performance factors…’), and so on. We had entire lectures on this stuff. And I don’t think the lecturer had been chatting to Dave Brailsford.
If this is the approach at an introductory level, imagine what it’s like when engineers are trying to be the best in their fields. And if everyone shares this philosophy, a more likely explanation for the success of British Cycling / Mercedes / Google is…money. It’s no coincidence, after all, that British Cycling’s dominance on the world stage has come in the wake of the introduction of National Lottery Funding for elite sport in 1997, and each medal haul has brought in ever more cash. Throw in a big sponsor like BSkyB (or HSBC) and you’re laughing. With money you can buy the best engineers and materials and doctors (ahem) and whatever else, allowing you to do the same stuff your competitors are doing, but more of it — and better.
The application of engineering’s attention to detail to other man-made systems — healthcare systems, for instance — is all very well and good, and will likely continue apace as data collection and analysis techniques become cheaper and more effective. But the marginal gains philosophy was supposed to be revolutionary primarily in its application to elite sport, and on its benefits for athletes Syed is now reticent. Perhaps he’s realised that here, too, the approach is nothing new. It’s what underpins sports science, which long predates Brailsford’s tenure at Team Sky — see here for more on that.
But even with sports science the gains to be made by individuals are few and far between. When it comes to the details of what ‘diet, training methods…[and] recovery protocols’ work best, the jury still seems to be out. Which should probably come as no surprise: biological systems are much more complex than anything humans have made, and we are still a long way from fully understanding them. We might agree in broad terms on what seems to work, but everyone knows what those things are. You put together talent, some consistent training, and a sensible approach to recovery that suits you, and see what happens. Which, as I understand it, was Wiggins’ point. Here’s hoping your competitors are too busy in the wind tunnel to realise.