Canadian Lawyer

18 J U L Y 2 0 1 7 w w w . C A N A D I A N L a w y e r m a g . c o m The Holy Grail, however, is not about reaching perfection in our predictions. It is about improving the ways we can come up with a better answer; to be bet- ter at predicting and more often. But prediction is hard. And it turns out humans are pretty rubbish at it, too. The inability to predict, for exam- ple, Brexit, the Leicester City Premier League win (an English football refer- ence — it was 5,000-1 odds!), or Trump winning has been blamed on many factors, from the fact that randomness prevails some of the time to our very real biases that mean we misjudge the odds of something happening. Being able to forecast what will hap- pen is why people approach experts in the first place, whether it's a patient wanting to know the risks involved in taking a certain course of antibiotics or a basketball scout in predicting who on the college courts will be the next great player. In study after study, though, prediction by algorithms outperform human judgments. This is leading to a growing trend to mistrust human intu- ition and rely instead on a more statisti- cal modelling process. The quote above is from Michael Lewis (author of Moneyball) and his new book, The Undoing Project. In it he reviews some of the heuristics and bias- es that Daniel Kahneman covered in his seminal book Thinking, Fast and Slow. It is these inherent human biases that affect an expert's ability to make accurate predictions. Lewis and Kahne- man explain that we tend to: • underestimate the occurrence of "Black Swan" events (Nassim Nicho- las Taleb's book, also called Black Swan, examines these events); • ignore the effect that sample size has on extrapolating conclusions — vari- ation from the mean is more likely in smaller sample sizes; • have an over-reliance on mental shortcuts — these shortcuts mean we get to avoid doing any maths or statis- tical modelling in our heads (which, as it turns out, is pretty important when predicting outcomes); • compare whatever we're judging to some model we have in our minds — which may be way off the reality sitting in front of us; • compare it to something that has recently happened or that was vivid and remarkable — called the "recen- cy" or availability bias. @k8simpson L E G A L I N N O VAT I O N N O W O P I N I O N Knowledge is literally prediction." Predicting outcomes better and more often is the great quest of recent times. Everyone is seeking the knowledge and insights that will make their predictions more accurate, from understanding the data points to antici- pating how the Houston Rockets will succeed in the NBA to every company that wants to know what you're doing on the web to predict what you might look for and buy in the future. Law departments want predictability, too. They want better prediction of costs and effort involved. In pre-litigation matters, clients want greater predictability in both the type and location of risk. Post-litigation, clients want greater predictability of outcomes to quantify their risk profile and in price predictability to manage costs. Predicting outcomes Humans use mental shortcuts in their daily lives, but expert predictions need to overcome this tendency By Kate Simpson "

• Cover