‘Chance dominates the game’ concluded C. Reep and B. Benjamin in their 1968 study ‘Skill and Chance in Association Football’ - and not without consequence. Until recently, this statement stood as received wisdom, the phrase deemed self-evident, its veracity left unquestioned. Predictions based on statistics were a folly they said, the game was ‘too fluid, too unpredictable’.

Although referencing football – specifically, the English variety - the sentiment was also thought applicable to sport more generally. Because eye-catching qualities like creativity and flair are difficult to quantify, the consensus was that judgement of sporting prowess was best left to those well-versed in said sport - in the majority of cases, former players. Certainly, the very notion that statisticians had anything to contribute in terms of team selection, acquisition, or even strategy was given little consideration.

Statistical Analysis, the Roots

In the UK, this began to change in the 1990s, when - as documented in the New Statesmen - undergrads at Lancaster University analysed, in-depth, goal-scoring probability in a hypothetical football match. While rudimentary in its technique, the analysis attracted the attention of more senior statisticians, namely Stuart Coles, who then expanded on the students' findings. His resultant work was to have far-reaching consequences.

Across the pond, a similar development was already underway. This one, however, was to be even more transformative, radically altering the way certain sports, particularly baseball, were conceived. Owing to its easily isolated ‘one-on-one’ contests - pitcher v batter, for example - baseball lends itself well to statistical analysis. Indeed, the use of sabermetrics, ‘the empirical statistical analysis of baseball’, has been employed in some capacity ever since the sport was first professionalised. Its current use, though, both its universality and sophisticated nature, has its roots in Billy Beane’s tenure as General Manager (GM) of the Oakland Athletics, and his now much-famed ‘Moneyball Effect’.

The 'Moneyball' Effect

As GM of the (relatively) cash-strapped Oakland A’s, Beane’s innovative approach was the perfect manifestation of necessity being the mother of invention. With no way to meet compete with the spending power of their deep-pocketed rivals, Beane adopted an analytical, ‘sabermetric’ approach when planning the assembly of his squad, bringing into his management team a number of statisticians for the purpose. Through rigorous and incredibly time-consuming analysis, the team learnt that ‘stolen bases’, ‘runs batted in’, and ‘batting averages’ – for years the yardsticks by which a player’s ability was measured - were not necessarily the best gauges of overall effectiveness. Rather, it was not-so-glamorous ‘on-base percentage’ and ‘slugging percentage’ that were more accurate indicators of players’ potential offensive success.

Historically, scouts overlooked these traits in favour of speed and contact, meaning their market price was undervalued, and players possessing said traits could thus be acquired on the cheap. It was following these new principles, heeding the word of his statisticians that Beane assembled his team. The result dismayed many, for on paper, the newly formed roster was a motley crew of old, in some cases injured, has-beens. Their appearance alone drew ire from the purists already angered by the cold scientific approach. How could a team of such visible misfits go onto success?

Well, initially they didn’t. The Athletics fared so poorly during the first part of the season - enduring at one stage a 5-16 streak - that many were quick to declare the method a failure. And yet, as many will know through its dramatisation in the film ‘Moneyball’, their fortunes reversed, the A’s eventually winning the American League West, breaking the long-standing record of 19 consecutive victories in the process.

The Spread

Around the world, these developments did not go unnoticed. The sporting community watched keenly as baseball laid down the statistical blueprint for success. It was not long before comparable models were incorporated into the management systems of football teams, both American and English, Ice Hockey, Basketball, and many others.

One of the highest profile converts was American football coach Bill Belichick. Although denouncing its use publically (to guard his methods), the New England Patriots boss had long held ‘a Moneyball sort of attitude’. In fact, the whole management team at the Patriots was imbued with a Moneyball-esque philosophy. From the owner Robert Kraft, described as the ‘biggest believer in data and analytics’, to Belichick's most trusted advisor Ernie Adams, the Patriots’ football research director and previous bonds trader, most were willing converts. Even Belichick himself holds a degree in economics. Such expertise, NBC has commented, is easily read into the team’s recruitment which is highly suggestive of a sophisticated understanding of ‘hyperbolic discounting’ - the result, thus far, has been the acquisition of high-quality players at bargain prices. On the field this has led to five Super Bowl triumphs - it seems to be working. Professional basketball teams were also quick to jump on the bandwagon, compiling huge databases of information from games that were then used in the offseason to inform decisions regarding contracts and free-agent signings.

Many a romantic will be turned off by the use of intensive, analytical thinking almost instinctively. The deployment of cold, emotionless figures to make sense of sport, something that by nature is highly dynamic, emotional, and capricious, just doesn't sit right. Such a mindset, though, ignores the creative latitude many of these models allow, particularly in more fluid games like English football and basketball, where a value judgement must still be made as to which parts of the game are prioritised.

But in the End...

Obviously, it will never be possible to reduce human behaviour to a simple formula - no matter how badly the betting enthusiast may wish. Yet, what analysts have proven, through the employment of statistical models, is their ability to get teams punching above their weight - a weight normally dictated by financial restrictions. That said, the advantage gained, history has shown, can often be short-lived. As the A’s experienced, the problem was that as soon as these models gained credibility, they were quickly adapted, and in some cases, bettered by other teams. Just two years after the A's experiment, the Boston Red Sox also adopted a sabermetrics model to aid recruitment. The result? They won the 2004 World Series, their first since 1918.