Sure way to win in only 500,000 years

IT'S random, addictive and, because it offers an expected pay- out of less than 50 per cent of your investment, it has been described as a tax on stupidity, yet a detailed analysis of the winning numbers and top prizes in the first 24 weeks of the National Lottery does reveal a strategy that can promise a profit.

The key is understanding the behaviour of other punters. Look at the first few lottery draws. Week one, back in November, produced the winning numbers 3, 5, 14, 22, 30 and 44. Seven lucky winners shared the £5.9m jackpot. Why so many? Aah, said the wise men from Camelot, it's because there were so many small numbers. People bet on their birthdays, you see. Week two produced 6, 12, 15, 16, 31 and 44 and four shared £7m. Birthdays again, they said. But week three gave us 11, 17, 21, 29, 30 and 40 - another five potential birthdays - and nobody won at all. It was the same on 7 January when 2, 5, 21, 22, 25, 32 produced not a single winner.

The following week, however, 133 people picked the winning sextet of 7, 17, 23, 32, 38, 42. Not many birthdays there. But the lotterologists were again ready with explanations: "It's people betting on their ages," said one pundit; "It's the effect of the lucky number 7," said others, drawing attention to the 7, 17 and 42 (6 times 7) among the winning numbers. But three weeks later we had only one winner when the draw produced the equally seven-rich set of 1, 7, 37, 38, 42, 46.

But if not birthdays or sevens, what guides people's choices? The clue lies in that remarkable week of 133 winners. What is so extraordinary about the set of 7, 17, 23, 32, 38, 42 is that there is nothing extraordinary about it. A large number of people clearly believe, irrationally, that an evenly spaced set of numbers is more likely to win than a closely clumped batch. Nobody, unless wilfully perverse, would bet on 1, 2, 3, 4, 5, 6. Somehow it looks much more unlikely than a well-scattered set.

That theory is confirmed by the result of 29 April when 9, 15, 22, 31, 34, 48 produced 14 winners sharing the jackpot. A good spread of numbers led to the second highest number of winners. And that provides the key to a winning strategy.

Look at the size of the jackpots in the roll-over weeks. On the four occasions when the top prize has not been won, the following week's jackpots have been £17.5m, £16.2m, £22.2m and £18m. If you can be reasonably sure that you will, if your numbers come up, not have to share the prize, then a flutter in such weeks is a good investment. You will still win, on average, only once in 14 million times, but your prize will be a healthy profit on the £14m invested.

The winning strategy is clear: pick a set of numbers so clumped together that no one else would choose them, and only buy a ticket when there was no winner the previous week. The only drawback is that, buying one ticket every six weeks (that's how often, on average, the jackpot is not won), and winning once in 14 million times, it will take over half-a- million years before youcan expect to be in profit. By which time, say astronomers, a meteor may well have crashed into the Earth, destroying all life. What rotten luck. Apart from that, the strategy's faultless.