Maybe simple statistics do trump judgement

Every time I read a review of 'Thinking Fast and Slow' by Daniel Kahnemann, a psychologist who won the Nobel prize for economics, I think I should read it. Especially in the light of a decision that is weighing heavily on me.

Most of the reviews, like this one, by esteemed physicist Freeman Dyson, start with a story, told in the book, in which simple statistics triumph over human judgement.

Invited to set up an interview system for the Israeli army, Kahnemann found that the existing system, relying on informal interviews, failed to predict which candidates were most suited to which jobs.

Inspired by a study that found predictions based on simple statistical scoring are generally more accurate than predictions based on expert judgement, Kahnemann replaced the interviews with a rigid questionnaire. The answers were collated and given numerical scores, and the candidates were matched more effectively to jobs.

Kahnemann had found his first cognitive illusion, the illusion of validity, which is a false belief in the validity of our judgements. And he includes examples where simple scoring systems are superior to the judgement of doctors and marriage guidance councillors, for example. The latter is particularly salacious, by the way, the Dawes formula that predicts marriage break-up, which is:

frequency of love-making minus frequency of quarrels

No doubt Kahnemann has studied the illusion of validity as it applies to finance. Many of his experiments involve the decisions we make about money.

The possibility that 'expert' judgement could be counterproductive bothers me, because I exercise judgement.

My approach to investing is essentially statistical. I screen the market for companies that pass statistical tests relating to value, and financial strength. Then I ask softer questions, perhaps akin to those Kahnemann might have asked of recruits; does the company face new and superior competition? Is the management team responsible for past success still there, or has the management team responsible for past failures gone? Are management incentivised to reward shareholders?

But then I do something interesting, that so far I have assumed adds to my chances of success. I let judgement take over and decide which factors are most important.

But I'm often tired by the time it comes to pass judgement on a company. I've computed a swathe of statistics, asked numerous questions, paged through annual reports. It's easy to let instinct take over, another theme of Kahnemann's book.

The judgement 'phase' is an open door to illusion. To be dazzled by one particular factor, the cheapness of a company, or the attractiveness of its product, and ignore others that together may be more important.

This tendency is worrying me because I'm thinking about the portfolio's big failure, Armour, and I think I made a misjudgement at that critical final stage. I'll save the details for the post-mortem, which is coming very soon, but I think, when I first added the company to the Thrifty 30 in April 2010 I allowed my need to add companies to the portfolio trump my doubts about the sustainability of past earnings.

In other words I was responding to another illusion, perhaps an even more deep-seated one: that it's safer being part of the crowd. The Thrifty 30 was young and half cash, and although I didn't recognise it at the time, I felt impelled to populate it in case the market rose and ran away from the portfolio.

I committed the original sin, going with the herd, which Kahnemann probably deals with extensively in his book.

I should read it.