The human brain is often described as the most complex system in the known universe. Predicting how it will behave should therefore be a huge challenge. In reality, it’s often a breeze.
Take the recent debacle over how hackers broke into the Apple accounts of celebrities and published their photos on the internet. Apparently a flaw in Apple’s iCloud service allowed a simple program to find out who was using predictable passwords. And it seems many celebrities were doing so.
Be wary of schadenfreude, however. According to William Poundstone, a former physicist turned best-selling author, such incidents are symptomatic of a fundamental flaw that virtually all of us possess: a feeble grasp of randomness.
In his latest book, How To Predict The Unpredictable, Mr Poundstone shows the ubiquity of this key human failing, and how it can be exploited.
Many games involve trying to outwit one’s opponent by behaving in supposedly unguessable ways – which, it turns out, are much more guessable than one might think.
In the game of “Rock, Paper, Scissors”, players like to think they’re good at randomly switching between the three gestures, making them hard to second guess. Research shows that most people are kidding themselves.
Most players tend to avoid “Scissors”, but for reasons perhaps best left to psychologists, male players have a particular preference for the fist-like “Rock”.
Knowledge of this lack of randomness shows that the optimal choice in any single play is “Paper”, which beats “Rock” and loses out only to “Scissors”. Of course, now that this little secret is out, it may not work as well as it used to. So Mr Poundstone outlines another trick: announce what gesture you’re going to choose – and then do it. Most people can’t believe you’ll eschew a random choice in favour of a true declaration of intent.
Some manifestations of such human predictability are rather more serious. The outcome of everything from driving tests to professional examinations has long depended on the use of multiple-choice tests.
They’re widely regarded as a simple way of distinguishing those who know from those who simply guess, with the correct answer being randomly buried among the various options.
Or at least they’re supposed to be. After all, it would be pretty disastrous if there were an exploitable pattern that know-nothings could exploit.
Yet Mr Poundstone has uncovered evidence that such patterns do exist, thanks to exam designers who aren’t very good at mimicking randomness.
After analysing the locations of the correct answers in tests for everything from college grades to online safety, Mr Poundstone found strategies that anyone stumped for answers can benefit from.
In simple true/false tests, one should guess “true”, as exam designers typically include about 25 per cent more of these responses than “false”.
On multiple-choice tests with four possible choices, the second one is most often correct, while on five-choice tests, the final choice is the best pick for guessers.
There’s a flip side to our inability to act randomly, however – and one that’s just as exploitable. While we like to think of ourselves as rational, most of us are easily fooled into seeing significance in meaningless noise.
Mr Poundstone calls this “hot hand theory”, after the belief – common among basketball fans – that players on high-scoring streaks are doing more than just being lucky.
In fact, such streaks occur through randomness far more often than people think. Given utterly random sequences of hits and misses to examine, people routinely see patterns that fool them into believing that a player possesses a “hot hand”.
The danger here is that this theory tempts us to believe that we can see the future.
If something is really responsible for a run of decent performance, then it makes sense to expect the past to be a guide to the future. But if the “hot hand” is an illusion caused by randomness, the past tells us precisely nothing about the future.
The problem, of course, is that it’s not always easy to decide which explanation is true. And as Mr Poundstone shows, there’s one notorious arena where both really do apply: the stock market.
Everyone from Nobel-prize-winning economists to amateur investors argues about the predictability of stock markets. A classic argument against predictability claims that the markets are “efficient”. That is, if there is any predictability, it’s quickly spotted by Wall Street’s finest, thoroughly exploited, and thus rendered useless.
On the other hand, a graph of the stock market in the United States over the last century shows a long, upward curve, which suggests predictability on some timescale. But what is it, and how can it be used?
According to Mr Poundstone, some answers have been found by Robert Shiller, a Yale economist and Nobel laureate. Along with fellow economist John Campbell, Mr Shiller has devised a measure of stock-market performance known as the Cyclically-Adjusted Price-Earnings (Cape) ratio.
While television pundits focus on market movements over periods of weeks, days or even hours, the Cape ratio ignores all this. Instead, it looks at market behaviour over the previous 10 years.
Only on that timescale can one see through the random jitters to glimpse the market’s “hot hand” behaviour. And according to Messrs Shiller and Campbell, this provides a pretty reliable guide to future market performance.
Stock-market data that harks back to the 1880s show that when the Cape ratio is markedly above its long-term average, the market is heading for a cooling-off period, when shares will perform relatively poorly.
Thus the time to buy shares is when the Cape ratio is lower than its long-term average of about 16.
Now for the bad news: the Cape ratio for the S&P 500 index currently stands at over 26, a level only previously surpassed in 1929, 1999 and 2007 – the dates of the three most notorious financial-market crashes.
Mr Shiller has stressed that he isn’t saying a crash is imminent. But we may be advised to brace ourselves for a long period of disappointing returns.
Whether the Cape ratio is as trustworthy during these strange economic times as it once was remains to be seen. What is certain is that most of us need all the help we can get to differentiate market signal from noise.
Robert Matthews is a visiting reader in science at Aston University, Birmingham.
newsdesk@thenational.ae
