Every Saturday, Danny write a sports piece for the Times called the Fink Tank where he looks at the probability of football match results and as he notes, the reaction to them is a little like the reaction people have had to the Met Office in the past week or so,
The most common e-mail tells me that I have incorrectly predicted a result. You said, the correspondent informs me, that my team would lose on Saturday and they went on to win 2-0. So there.Amen to that brother I say. This is why whenever I look at the weather report I tend to use the Weather Channel website and, instead of just looking at the logo for the weather on a given day, I look at the little bit below that shows the the percentage probability of precipitation.
I reply, as patiently as I can, that I didn’t forecast defeat. I merely said that it was more likely than victory. There is, I emphasise, a big difference.
The reason why we provide match probabilities rather than forecasting the result is that forecasting the result is, ahem, impossible.
If I could forecast the result, I say, I would be rich beyond my dreams (true) and wouldn’t be sitting here typing articles about football (a lie).
And the same is true of the weather. People watch the weather forecasts to be told whether it will rain or shine on some future date.
But providing this information is impossible. The best the forecasters can do is to say what the weather will probably be like and how high this probability is.
I’ve long been highly critical of the standard weather forecast because it leaves out this critical information.
See how tonight they say there is really only a 30% chance of rain, ergo a 70% chance of no rain at all.
Why, exactly, the BBC weather reports on TV cannot show this information I do not know. I have noticed in recent weeks that the icon is usually a black cloud, with rain and sun on it, which I often joke means "we haven't a bloody clue".
Perhaps the BBC and other TV weather forecasts could start putting a little % value under their icons on the maps, then at least people would start to realise that weather forecasting is about probability of weather, not about what is going to happen