It’s a fact of statistical life that random events tend to form clusters. If you think about it, it’s extremely unlikely that unconnected happenings will continue to be evenly spaced with clockwork regularity – so clusters provide the only alternative. Yet our pattern-seeking brains have huge trouble with this. We see a cluster of events – either close in time, or close in space – and assume there has to be a common cause. (Of course there could be, but often there isn’t.) I have now officially decided I hate clusters.

Two examples, the second of which triggered this irritation.

I sell copies of my biography of Roger Bacon The First Scientist via Amazon Marketplace. On average, I sell one every couple of months. But a while ago three sold in the same day. This seemed extreme enough to suggest a cause and I emailed the buyers to find out why they’d bought it – but no, it was a purely random cluster. (A little while later, as if to rub it in, two sold within an hour of each other.)

Last night, though, I saw clustering with a vengeance. I was tired having had one of those holidays that’s more exhausting than being at home. I was looking forward to a good night’s sleep, with the extra hour afforded by the end of daylight saving time. Only at 3.15am one member of the household woke me up, needing some painkillers. At 4.30am another member of the household woke me up because she had an upset stomach. At 6am the dog, apparently not aware it was now GMT, decided it was time to be let out. And at 7.40am an alarm clock that had become accidently set when removing it from the suitcase went off.

So, happy though I am that there was no evil fate involved, just good old statistics, I still feel that all clusters should be taken out and shot at dawn.