I’m not usually interested in Maths but…

I was never very good at Maths when I was at school, although thanks to the excellent teaching I received, especially at primary school, I am quite good at mental Maths… but Maths as a subject leaves me cold, and I just haven’t the brain for the abstract understanding needed.

However… I was listening to a programme on the radio as I was driving, and it was about Maths, but Maths in relation to actual events.

The programme started by mentioning the fact that there has been a spate of train crashes across the world recently, in Canada, Pakistan, France, Spain and Switzerland. Inside Science, the programme I was listening to, was discussing whether it was a trend or just a coincidence. The presenter asked Professor David Spiegelhalter of Cambridge University and an expert in the public perception of risk, to explain whether there is such a thing as a ‘crash season’, whether a cluster of events revealed an underlying problem or whether it was just bad luck and chance.

Professor Spiegelhalter answered that there was no common pattern in the recent train crashes, it was just chance and “chance works in a strange way.” On average there are about fifty major train crashes a year, so that averages at one a week, but obviously that doesn’t really happen and there can be clusters of a couple and then a couple of weeks with no such events. So far this year there have been twenty-nine train crashes, so that is about right as far as an average goes… by  mathematics there should be some weeks in a year where there would be three train crashes.

The Professor spoke of the recent horrific and tragic Spanish train crash, and said that although there might be factors regarding the driver, when the way the new track was built was considered there were other aspects, for example the fact that the railway line was very straight  and then there was a sharp bend and  a concrete wall adjacent to it,

The professor then spoke about the statistics for rail safety in the UK; we are extremely safe and there has been no fatality caused by a train crash for five years; a person would have to travel 7,500 million miles before there would be a risk of them being killed in a train crash. About 300 people are killed by trains every year, sadly most of them are suicides, but about fifty are accidents.

He related an interesting piece of information about  flying; following the dreadful events of 9/11, many people stopped flying and began to drive. As a result there were 1500 deaths on the roads from people transferring from flying to driving.

The way events are reported can also give the impression that there is a spate of a particular sort of accident; for example, recently in Britain a lot of people have drowned while swimming. More people went swimming because of the hot weather, and a lot of those people were swimming in dangerous places such as quarries and rivers. If a single person had tragically drowned in this way it may have been reported locally but not more widely; because there were a lot of events, even though they were explainable, they were reported as a cluster, the underlying risks not completely  appreciated in the report.

Similarly, about three years ago there were four murders on one day in London; this was reported as being the start of a crime wave, and the deaths were attributed to many things. In fact, by looking at the statistics, it was clear that there are normally about 3 murders a week in London, but obviously it would  be as regular as that, so there maybe more one week, and then fewer for a couple of weeks. Overall the murder rate in London is going down.

People (i.e. us, the general public)  are bad at recognizing randomness, we see patterns where there aren’t any; stories suggest patterns and then as with several train crashes, or murders, or drownings, the facts have to be examined and the idea of there being a pattern knocked down…. so many apparent patterns are really just bad luck or coincidences.

 http://www.bbc.co.uk/programmes/b037jgll