What Is A Forecast?

Just following up on my earlier post of the dissembling and games being played with hurricane "forecasts" by insurance industry lapdog RMS. Basically, Roger Pielke asked RMS if they are standing by their forecast of a 40% increase on insured losses due to hurricanes for the time period 2006-2010, after both 2006 and 2007 resulted in below average losses. From Prometheus:

Yes, we do stand by that forecast, although I should point out that we update the forecast every year, so the 2005 forecast (for 2006-2010) is now 2 years out of date. Apart from questions of forecast accuracy, there's no particular reason for any of our users to use the 2005 forecast at this point (that would be like using a weather forecast from last week).

Think about this for a second. The RMS "forecast" was used by insurance companies as an excuse to raise the rates consumers pay for coverage, because they called for a 40% increase in damages. The original RMS forecast has, to this point, been off by a factor of 20. Actually losses have been 95% less than RMS forecast. But RMS claims, that's okay, because they now have a new forecast which raises losses by another 40%! Based on what exactly? Because the old forecasts have proven so inaccurate?????

The "logic" presented is that we need not evaluate the old forecasts because new forecasts are available. What garbage. You can only evaluate any methodology based upon past results. You make a prediction and then you compare that prediction to actual events. Period. But, using RMS' rules, their methodology can never be evaluated because they will always be issuing new "forecasts." Baloney. A forecast for the period 2006-2010 is not "out of date," because we haven't reached 2010 yet. Period. A new forecast for the dates 2008-2012 doesn't take the place of the older forecast because it covers a different period of time. It is a forecast of something else.

RMS' response wasn't complete nonsense, however:

One curious property of the loss distribution is that it is very skewed. As a result the real losses would be expected to fall below the mean in most years. This is compensated for in the average by occasional years with very high losses.

Yes, that is obviously the case. Two years of low loss totals does not mean you can say there is a "trend" towards less loss occurring generally, any more than a Katrina type event can allow you to make the opposite conclusion. The point is RMS has to be committed to the 2006-2010 forecast win, lose or draw. It seems unlikely that they will be able to make the case for their 40% increase, barring another Katrina. If their "forecast" fails and RMS and the insurance industry don't take that into account, then they are engaged in nothing but an attempt to bilk their customers.