If you had turned up to any match played by the then Division Three (South) side Swindon Town FC in 1950, the man who was arguably the father of football statistical analysis would have been fairly easy to spot in the crowd. Charles Reep was the one meticulously sketching out play-by-play diagrams of everything that happened on the pitch – and occasionally sporting a miner’s helmet to help him to see his notes.
OK – so it may not have quite the glamour of Brad Pitt playing the number-crunching baseball coach Billy Beane in Moneyball but as this article on the US-based FiveThirtyEight website points out: “More than 60 years before player-tracking cameras became all the rage in pro sports, Reep was mapping out primitive spatial data the old-fashioned way – by hand.”
What Reep came to spot from his reams of notes and diagrams was that most goals in football resulted from ‘plays’ that were preceded by three passes or fewer – from which, to his way of thinking, there was one obvious conclusion. Teams should concentrate on getting the ball down towards the opposing team’s goal as quickly as possible – and thus was born ‘long-ball’ football.
It took a few decades for Reep’s ideas to be picked up in the higher reaches of the game but, in the 1980s, the long-ball style of play famously helped Wimbledon move from the fourth to the first rung of English football, not to mention an unlikely FA Cup win in 1988. It was even adopted by the England team during Graham Taylor’s stint as manager from 1990 to 1994 – albeit with rather less success (and did we not like that).
The trouble was, as FiveThirtyEight puts it, Reep’s theory was based on “a fatally flawed premise”. His mistake was to focus on one specific area – the percentage of goals generated by passing sequences of various lengths – but this fails to acknowledge that football is a game where possession changes a lot. Ultimately, retaining possession gives a team a better chance of scoring goals than booting the ball miles up the pitch.
In developing his theories on long-ball football, Reep fell foul of a number of cognitive biases that can often lead investors astray. As Daniel Kahneman argues in his book Thinking fast and slow, our brains are inclined to ignore what we cannot see and to give greater weight to what is in front of us – analysing only that data, reaching a conclusion on it and then fitting a story around it.
Not only can this see people becoming overconfident on the basis of what is in fact a small set of data, it can also lead to the behavioural sin of ‘base rate neglect’. This has nothing to do with interest rates but involves attributing less weight to more general probabilities (in Reep’s case, greater possession leads to a greater chance of scoring a goal) and more to newer or more immediate information (length of passing sequences).
A good example of this happening in the world of investing is the way people can often think about growth stocks. Investors tend to look at businesses that are currently small but fast-growing and what immediately comes to mind are not the significant number of such companies that went bust or whose growth quickly slowed but the high-profile yet comparatively rare successes such as, say, Amazon and Google.
What makes value investing such a tricky discipline for most to adhere to is that it is about understanding base rates and investing not on the basis of what is right in front of us – a couple of years of good performance, for example, or limited amounts of information – but according to which wider probabilities are in our favour across the largest sets of data we can find.
Here on The Value Perspective, we believe assessing these wider probabilities and buying cheap but financially strong businesses is the equivalent of possession-based football because more than a century of history has shown, on average, it works – and, again on average, puts us in a stronger position than those investors who prefer, as it were, to hoof the ball up the field and hope for the best.