I just read a superb book called ‘Fooled by Randomness’ by Nassim Nicholas Taleb. I wish I had read it before, and can see that it is a book I will re-read a couple of times. It would be trite to try and summarise what the various premises are, but a rough cut is that it it concerns the way in which people fool themselves that they understand the cause and effect in things that are actually random, and worse that they understand the bounds of their knowledge. Though it is well founded, it is far from a dry academic read, with loads of real-world examples.
A couple of points stood out for me – not really that I hadn’t thought of them, just that they chimed well with my own prejudices. The first is what Nassim calls ‘Wittgenstein’s ruler’ which concerns the accuracy of measurement when the thing you measure with is of unknown accuracy. Particularly, it is likely that you find out more about the thing you are measuring with than the thing you are measuring. The example he uses is book reviews on places like Amazon, where you learn very little about the book, but a lot about the review writer. I think you see this even more clearly on comments on the internet – for example, Stephen Fry now has a blog (it’s here). It’s well worth reading, but if you read the comments, it is startling how little they inform, but just how much they tell you about the opinions of the writers, as well as their desire to look erudite.
The second point that chimed especially well with me was the excess confidence in cause and effect that people have … and particularly the extrapolation and overconfidence that they are then prey to. Economists and journalists come in for a particular slating over this. I was entertained that whilst I was reading this I had the news on, and heard the anchor discussing Labour party conference with the live reporter. Some comment or another had been made about Gordon Brown and David Cameron. The anchor asked ‘whether it is possible to calculate how much this has damaged Cameron’. No eye was batted over the impossibility of calculating it in any meaningful way. I have seen it many times in business, with wildly accurate predictions made on small tests, or even numbers pulled out of the air based on ‘experience’. Sometimes this is necessary (before anyone says that they have seen me do it) – but you need to retain a high degree of scepticism and I have routinely seen this not done, and then surprise expressed at an unexpected outcome.
He has written another book about black Swan events that I think I’ll read as well. Black swans are things that no-one has seen, but whose non-existence cannot be proved, and that seem to turn up more often than one might think. The name comes from an observation that no amount of seeing only white swans can prove that there can’t be a black swan … but one black swan proves that they can. A good example would be the total drying up of liquidity in the banking market. Never seen in a modern market, but that is far from proof that it can’t happen, as we have seen.
I could go on at some length, since I loved the book – instead I’d just say you should read it, since Wittgenstein’s ruler says that this note really tells you much more about me than the book 😉
One last part – I was again reminded why I HATE graphs that don’t start at zero. People are more sensitive to movements than absolute levels, so using graphs that pander to this flaw feels a bit cavalier.