Research interests My research interests extend from cosmology to the origin of urban myths, but a theme running through much of my work is the application of mathematics to questions often thought beyond the reach of detailed analysis. For example: how do we capture the idea of plausibility in assessments of some new supposed health risk ? Is there any good evidence for "paranormal" phenomena ? When should we put our trust in predictions of, say, earthquakes, and when should we ignore them ? It's a pretty vast field, and pretty under-explored one, but it often turns out that the truth is much more interesting that one might believe..... For a list of my research papers,  click here. Recent projects include: An easy-to-use method for assessing the plausibility of statistical findings (click here to use the technique to assess the plausibility of new clinical trial findings); The use and abuse of subjectivity in scientific research, and its role in a major scandal whose impact on the reliability of research findings across the entire scientific spectrum is still not fully appreciated; Bayesian analysis of evidence for anomalous phenomena, such as ESP and homoeopathy; Investigating whether a bizarre cosmic "coincidence" first noted by Paul Dirac in the 1930s can explain a number of major problems in cosmology and quantum field theory; Using Decision Theory to analyse the vexed question of whether it will ever be possible to predict major earthquakes. The polite mathematical version is available here . The bar- room version is: "Stop wasting everyone's time and money, guys - you can't do it". The results have interesting implications for deciding when to take an umbrella following a forecast of rain . Applying Bayesian inference theory to the analysis of legal evidence such as confessions (which turn out to be especially unreliable in terrorist cases) Using probability theory to explore the reasons why people reach for paranormal explanations of "spooky" coincidences and to predict real-life coincidences The training of neural networks to recognise literary styles , and the use of such networks to cast light on literary mysteries such as Marlowe's influence on the young William Shakespeare Extracting a (surprisingly accurate) value for "pi" from the appearance of the night sky Analysis of urban myths, such as manifestations of Murphy's Law - "If it can go wrong, it will". Published results cover the notorious examples of tumbling toast landing butter- side down, why there are so many odd socks in our drawers, why rope, string or flex so often seems to acquire knots , and why places you're looking for so often lie in awkward places on maps . My research here has led to a cover feature in Scientific American (April 1997), a Discourse to the Royal Institution of Great Britain , a feature in Reader's Digest (March 1998; I must tell mum), and various awards.