Just Finished Reading: The Physics of Wall Street – A Brief History of Predicting the Unpredictable by James Owen Weatherall (FP: 2013) [243pp] 

For as long as there has been gambling (a VERY long time) there have been those who tried to ‘play’ the system in such a way that they could ensure winning much more often than losing. For most of that history such actors have operated by using their instincts, magical thinking and, where possible, cheating or some sort of manipulation. As you might expect the vast majority of these schemes fall at the first turn of a card or very shortly afterwards. That all changed with the advent of mathematics and the development of probability theory. Once games of ‘chance’ and skill, such as Blackjack (21) and Roulette, became capable of being modelled mathematically it became possible to predict outcomes with a fair degree of accuracy – and certainly more accurately that the average player was capable of. But even with maths, theory and the odd (early) computer behind them these theorists found that making a LOT of money needed a LOT of initial investment, something which most of them and most of their university departments either didn’t have or were loath to loan either with a high probability of a low return. But when modelling card games started to fall out of favour the mathematicians and physicists saw a much more lucrative and much more challenging prospect – the Stock Market. 

Of course, modelling the Stock Market was never going to be easy. Indeed, many scoffed at the very idea of attempting to predict the unpredictable but, ever up for an impossible challenge, some mathematicians tried and, more radically, put their own money where their mouths were. Early models (and it needs to be kept in mind that these were models and not reality) of early Stock Markets – notably in France – where neither particularly accurate nor particularly responsive to sudden changes. But they didn’t need to be – at least generally. Even with the advent of the telegraph the amount of stock being traded was small and the speed of the trades almost glacial compared to today. Even so the early crude models had a long way to go and they only really began to take off with the advent of computers post-WW2. As computers got faster the capability of the Stock Market models improved in step – but so did the size and complexity of the Market itself. Still, many thought the task was simply impossible. There was no predictable market to model, they were wasting everyone's time and, more importantly, money. It was time to put up or shut up and that’s exactly what one team of physicists did by creating a Hedge Fund based on their theories. What happened next astounded everyone and changed financial management forever. Not only did the Hedge Fund consistently make money but it consistently outperformed other Hedge Funds made up of supposed market experts. The Quants had arrived.  

This was an interesting story, well told. It seemed obvious to me that Stock Market prices cannot possibly be random and that the ebb and flow of the Market cannot, therefore, be completely unpredictable. If that was the case there would be no way to regularly make money on the market as every deal would be a crap shoot. As at least some people obviously make a great deal of money from market trading, and do so consistently, there are rules governing the system that can be discovered, followed and exploited. Likewise, the better your model is of actual behaviour then the better and more consistent your results will be even when markets become volatile. What was more interesting from my PoV is that modelling the Stock Market is, essentially, modelling aggregate human behaviour. This mean that, at least theoretically and at least crudely, human behaviour – in the aggregate can be modelled mathematically if you have the right algorithms to do so. With my long-standing interest in Asmovian Psycho-History it raised the possibility that these early predictive tools for Stock Market trades could be the foundational (pun intended!) building blocks of such a theory. Definitely a recommended read for anyone interested in the Stock Market, mathematics or the possibility of prediction human behaviour. More to come on this and associated subjects.