One of the ways people comfort themselves after natural disasters such as the Asian tsunami is with the thought that, as such events are rare, it will be a long time before there is another one. But this is more superstition than science.

The fact that such a disaster happened last month in no way alters the odds of a similar event occurring next month, or even tomorrow. Probably not in exactly the same location, as far as the epicentre is concerned, but quite possibly in the same part of the world. The statistics that apply to such events are essentially the same as the statistics that apply to tossing a perfectly balanced coin - if it comes up heads on one toss, the odds are still one in two that it will come up heads next time as well. The only good news is that the odds of a disaster on the same scale are much smaller than one in two.

There are lots of ways in which you might guess that earthquakes (the cause of tsunamis like the recent one) are distributed in time. At one extreme, most earthquakes might be very large, releasing lots of energy which then takes a lot of time to accumulate again. At the other, most earthquakes might be very small, repeatedly releasing tiny amounts of energy and never doing much harm. Or there could be some typical size for an earthquake, with both larger and smaller events relatively rare.

Clearly, guessing is futile, and we need a proper scientific assessment of the statistics. Indeed, when the records of earthquakes are investigated to see how many of each size have occurred around the world over a long period, they show none of these patterns.

The first person to analyse the records in this way was Charles Richter, who introduced the eponymous scale used to measure the intensity of earthquakes. This scale is another source of confusion to the uninitiated. How come a magnitude nine event is so much worse than a magnitude seven event? The answer is that the scale is logarithmic. A magnitude two event is not twice as powerful as a magnitude one event but 30 times as powerful. A magnitude three event is 30 times as powerful as a magnitude two event, and so on. So a magnitude nine event is 900 times as powerful as a magnitude seven event.

Almost fifty years ago, Richter and his colleague Beno Gutenberg used this scale as a basis for investigating the frequency of earthquakes of different sizes. They combined quakes into "bins" of half a magnitude on the scale, so all the earthquakes between five and 5.5 went in to one bin, all those with magnitudes between 5.5 and six into the next bin, and so on. Since the Richter scale is logarithmic, they took the logarithm of the number of earthquakes in each bin, and plotted the results on a graph known as a "log-log" plot. The results lay on a straight line. Every subsequent study has shown the same thing.

This is one of the simplest (and most common) patterns of behaviour found in nature. It is called a power law, and in this case it means that for every 1,000 earthquakes of magnitude five there are roughly 100 earthquakes of magnitude six, 10 of magnitude seven, and so on. This law applies across a huge range - a magnitude one event is about the same as the rumble you feel when a heavy lorry passes your house, while a magnitude nine event is some 500bn times more energetic. The power law tells us that although large earthquakes occur more rarely than small earthquakes, they are produced by the same physical process. You do not need to invoke a special cause for why large earthquakes happen - they just do. The power law also tells us that for any magnitude you choose, there is a particular probability of an earthquake that size occurring during, say, any one year. Small earthquakes are relatively common, but they occur in the same random way as the way the numbers come up on a set of dice. Large earthquakes are rare, but they also occur at random. An earthquake of any size could happen at any time in an earthquake zone, even if one the same size happened last month.

The moral is plain. We should be no less, and no more, concerned about a tsunami disaster affecting the Indian Ocean today than we were on Christmas Day. If hindsight suggests that it would have been a good idea to have a tsunami warning system then, that need is exactly the same now. All this will have been obvious to anyone with A-level physics; such a pity, then, that only about 30,000 students took A-level physics in Britain in 2003, and that so few politicians have any background in science.

**·** John Gribbin is a visiting fellow in astronomy at the University of Sussex and author of Deep Simplicity (Penguin)