As mentioned in yesterday’s NNAHT, Bobby McMahon wrote a short piece at Forbes on Charles Reep as a forefather of sorts to the modern soccer analytics movement. McMahon writes of Man City’s performance analyst Gavin Flieg and the recent Opta data dump:
On the face of it, Fleig and City’s intentions are laudable given the huge increase in statistically based soccer articles and blogs. A majority of these authors must be aware of Bill James and no doubt, a good number have been been influenced and inspired by James’ work.
In the absence of a soccer pioneer then James seems to be a plausible stand-in.
Except for one thing – Bill James work with Baseball statistics was pre-dated by around two decades by a man the Journal of Sports Sciences called the first “professional performance analyst of football” after his death aged 98 in 2002.
So why wouldn’t Gavin Fleig give a nod to wing commander Charles Reep?
McMahon goes on to answer his own question: Reep, unlike James, was undeniably a quack. And this is one of the major problems with the famous RAF Wing Commander turned amateur soccer statistician: he was indeed a pioneer, a bona fide “performance analyst,” influencer of Charles Hughes, Graham Taylor, even Sven Goran Eriksson. He was also fundamentally misguided, and arguably helped set English football back a generation, if not more, whilst sowing the seeds of media skepticism surrounding football analytics that remain to this day.
What did Reep believe exactly? Essentially, that most goals came from moves involving three or fewer passes, in addition to opposition mistakes. Therefore, the best thing was for defenders and midfielders to send the ball over the top to one or two lone strikers, who would hopefully be in advanced enough position to catch opposition defenders out of position. Also known as, loooooong ball, route one, you need a big, strong centre forward, put it in the mixer, get rid etc. etc.
As McMahon points out, this doesn’t hold up to scrutiny, particularly as 90% of all moves in football go no further than three moves. But the essential problem with Reep, as Barney Ronay masterfully detailed in a 2003 When Saturday Comes column, was that despite his experience as an accountant, his approach to understanding football was marked more by ideology than science. Ronay writes:
The urge to debunk footballing complexities was a force behind much of Reep’s work. Among his published articles are the jaunty “This Pattern-Weaving Talk Is All Bunk!” (1961), the po-faced “Skill And Chance In Association Football” (1968) and the downright aggrieved “Are We Getting Too Clever?” (1962). In this article Reep goes on to recount his first meeting with Wolves manager Stan Cullis, shortly after the Hungarian demolition in 1953. Cullis, perhaps impressed by Reep’s work in helping Brentford avoid relegation in the spring of 1952, requested his assistance in devising a style of play that would borrow from the Magyars while reaffirming the “wholly English” principles of “direct passing”.
If bad first impressions account for anything, Reep’s abuse of rudimentary statistics muddied the water for football analytics in England for years to come. Beyond his love of long-ball against rhyme or reason, Reep came to a single, extraordinarily flawed supposition based on a bare modicum of evidence, and pedaled it as God’s Truth for the next forty years, and indeed the one and only system in which long-term success could be achieved in football.
While many in the current analytics field are quick to congratulate themselves on avoiding similar pitfalls in their current work, some of the same spirit of Reepism remains. Some would-be analysts reverse engineer their conclusions from the most limited of data in order to prop up whichever form of football they favour, including Spain and Barcelona’s tiki-taka approach. Some confuse the nominalism of their preference with the realism of the soccer itself, a sport in which the “winning” approach can take different forms depending on the skill of the players, the skill and tactics of the opposition, and the in-vogue style of the day.
And while there is certainly evidence that, over the long-term, certain metrics do positively coincide with a winning record over and above the reversion to the mean, these metrics often have little bearing on individual, ninety minute matches.
Football doesn’t trade in absolutes. Neither do most reliable metrics in football analytics. The data—and we’re only barely out of the data-gathering stage in soccer analytics—is still in many cases obscure, and an effective approach in one league across a single season cannot be used as a yardstick for the “way to play the game.” Neither should those who purport to study the game with disinterest, lest they repeat the same mistakes as Reep.
(Video of Reep-acolyte and Watford manager Graham Taylor on his team’s “style.”)