MB & DS’s Norm Chronicles
Michael Blastland & David Spiegelhalter The Norm Chronicles: Stories and Numbers about danger Profile Books 2013
Both authors have contributed to the public understanding of risk, with Michael coming from literature and journalism, David from statistics. The book uses three characters, Norm (risk neutral), Prudence (risk averse) and Kelvin (risk seeking) to persuade us, with stories and numbers, that – contrary to what some might suppose – that the conventional norm of risk neutrality is not necessarily a good guide to life. David also has a blog.
[Exactly] the same behaviour on our part can mean extreme differences of risk, depending on what other people do. This makes reliable numbers about the risks that you face as an individual impossible to calculate. There’s a big risk down the road for all of us if there is a failure of herd immunity, but what that means for you, now, is impossible to say. You could be fine. But what if you contribute to a loss of herd immunity? Youc could be dead. So what’s the risk?
[People] tend to assimilate new knowledge in a manner that confirms their emotional and cultural predispositions.
[The notion of a ‘MicroLife’, a reduction in life expectancy, is introduced to complement DS’s previous notion of ‘MicroMort’, the chance of dying now.]
Suddenly, chronic risk feels a lot more here and now.
… Kingsley Amis said ‘No pleasure is worth giving up for the sake of two more years in a geriatric home Weston-Super-Mare.’
… We can’t foresee everything that might go wrong, so we take the easy route of assuming it won’t. The remedy isn’t just better numbers, it’s more varied stories that can take us imaginatively out of our comfort zone – one reason why people are attracted to the idea of scenario planning as a way of thinking what future dangers might be.
… We think that there is more to uncertainty than you’d think from the way that people throw numbers around.
We think this especially because when you try to grab hold of probability it somehow slips through your fingers. It’s hard to say what it really is for an individual. Its hard too to say how the average affects any individual.
[Any] number that we claim for probability is based on what we know. It is necessarily a judgement and does not exist as a property of the outside world. Risk, in this sense, is a measure of what we don’t and can’t know as much as of what we can.
All of which forms part of a rather startling conclusion: that independent, objective probability, as Norm says at the last, doesn’t exist.
Nor, as we say, does the average person exist to whom the average risk is supposed to apply. The average is an abstraction. The reality is variation.
So in practical terms, for the events of life in general, when we say a certain activity is dangerous and quote its risks … these numbers should be considered only as reasonable betting odds given what we know. As soon as we know more … the risk changes, suggesting that the potential degree of refinement is often infinite. And this can easily be applied to things that have [already] happened but we don’t know about yet … .
- An interesting book, but I couldn’t help comparing it with the theory. Maybe it should have a geek annex.
- Some of the book’s points can be seen in terms of short-run versus long-run probability. Probability seems less problematic in the short-run, and it seems reasonable to apply it the longer term, provided that there are no life-changing events or other significant changes to context.
For vaccination, it seems reasonable to think in terms of the probability of catching a disease if you don’t get vaccinated in the short run. But if the take-up of vaccination should change there might still be – in some sense – a normative probability, but it might well be impossible to calculate. On the other hand, if your decision could influence others, then unless you regard yourself as a probabilistic automaton, the most you could do is to estimate probabilities conditioned on your decision, and even this seems doubtful.
MicroLives seem to me to be a good innovation. Sometimes when we reduce our life expectancy we may be running the risk of sudden death in later life, and hence losing years in Amis’s geriatric home. At other times we may reduce our life expectancy through premature aging, so we are losing our best years. It is also possible that we might age quicker. Which is the case should make a difference to the perceived ‘cost’ of the life-reducing activity.
The converse case is where exercise (or other healthy activity) prolongs our lives. If you don’t enjoy walking and you need an hour’s walk to prolong your life by 10 minutes (depending, e.g., on how fast you walk) then your enjoyable life is reduced: it doesn’t seem worth it. Similarly, if you think that some activity increases the ‘worth’ of each calendar year by 20%, you may not mind aging 10% faster.
Big Risks and Unemployment
- The book’s remarks relate to the difficulty of conventional probability theory that the probabilities must add to 1. A way around this is to work with odds (e.g. P(Heads)/P(Tails)). One can then identify which recognized possibility is most probable, leaving open the possibility of ‘something else’.
- All sorts of devices may be needed to overcome any prejudice against unusual possibilities. Some commentators think that as scenario planning became more widely adopted it became less effective at this. My own view is that appropriate mathematical modelling can be used to identify abstract potential equilibria, forming the basis for story-telling. Being abstract, the mathematics is less constrained by prejudice and catalyses challenging stories.
- A common problem is where a system (such as ‘the economy’) has settled into an apparent equilibrium. Many analysts will use statistics to make predictions, implicitly assuming that the equilibrium will continue. One particular needs stories (as in scenario planning) and suitable mathematics to consider ways in which the equilibrium may end.
- I might quibble about the use of the term ‘refinement’. The probability might change dramatically with increased knowledge.
- What is said about ‘any number that we claim for probability ..’ applies to what others say. Thus when someone says ‘your probability’ or ‘the probability for someone like you’ they really mean the probability for people that they happen to classify in the same way as you. Not only might your estimate of your probability be different because you are ‘an individual’ or because you have additional information (such as family history) but also because you classify yourself differently. For example, a recent immigrant to the UK from Poland might regard the Polish health statistics as more relevant to them.
The book shows that simple numeric probability does not always adequately capture uncertainty. The critique mirrors that of Keynes’ Treatise, but where Keynes, Turing and Good develop mathematics it uses stories. It seems to me that both mathematics and stories are needed, and that mathematics may be the best tool to explore novel possibilities, while stories can be used to check, explain and present the results.
David’s Don’t know Can’t know.