Oh, dear. Interviewers, particularly with Jeremy Paxman’s prestige, presenter of Newsnight and University Challenge, should know something about their subject before an interview. Paxman, interviewing Robert Chote of the Office of Budget Responsibility (the independent body evaluating the government’s forecasts) on the subject of government forecasts, displayed a level of ignorance combined with his usual aggression that was 'stomach churning'. At least to me as a forecaster.

He seemed to believe that forecasts were reasonably expected to be correct, while even a passing familiarity with the topic would suggest that such a goal is way outside our aspirations never mind our abilities. Robert Chote was pleasantly polite in correcting him and making the key point about understanding uncertainty. Accuracy is important of course but even more so is to estimate how wrong you can expect to be (and these prediction intervals too need to be accurate). Where Jeremy Paxman could usefully have criticized the OBR is that its estimates of uncertainty were also wrong! And come on, Jeremy, can you demonstrate some level of understanding rather than playing to the lowest common denominator in the audience?

Professor Robert Fildes