In a 2009 paper on the ability (actually, lack thereof) to predict US hurricane landfalls or damage on time scales of 5 years or less, I illustrated the perils of correlation shopping with the data shown in the graph above (paper here in PDF).
Over the period 1950 to 2007, in years which the score of the FA Cup final totaled 3 goals or greater, US hurricane damage in the subsequent season (June through December of the same year) was 31% greater than average. In years which saw 2 goals or less scored in the final, hurricane damage was 33% less than average. The relationship is remarkable. Chance you say? I once thought so too.
Little did I know that I had stumbled on to a major scientific breakthrough. Since 2007, the FA Cup hurricane damage predictor (FACHDP for short) has correctly predicted 3 of the 5 seasons damage correctly. By contrast, sophisticated catastrophe models used in the financial world to assess risk correctly anticipated only 2 of those 5 years. Obviously, they have not yet caught up with the implications of this remarkable new discovery.
The mechanisms underlying the surprising predictive ability of the FA Cup are still being explored, but they are no doubt reflective of what some have called "a new normal" in hurricane behavior.
US coastal residents, Bermuda reinsurers and City financiers await today's match with great anticipation.
No comments:
Post a Comment