Tuesday 18 November, 2008 An email from a reader this weekend quite rightly accused me of forgetting one of the key lessons of modern analysis, be it scientific or economic, namely that past observed relationships don’t always hold true in the future.

My correspondent was taking me to task for detailing the research by the Directors Dealing Trust that started in 1994 – in the past there may have been a relationship between strategies that focus on these purchases but that it is no guide to the future was the ‘big point’ I had failed to mention. To be fair I frequently warn that back testing is a dangerous business when it comes to constructing investing strategies but the point is well taken – the ultimate test will be over the next year as the fund managers put their analysis into action.