We have a new month and new data
for the dynamic trading risk factor. Last month's forecast was for a profit of 2.17%; the
realization was a profit of 4.33%. This month we are forecasting a profit of 2.11% for June, 2009.
(Note that the underlying charts for the prior posts have been replaced by the newer versions.)
In order that we make the best possible forecast of the next months return, the data linked
to above contains forecasts that computed from regression parameter estimates that have been updated to take account
of the increased data set available at the time that the series is computed. On this basis, only the final forecast
in the series is truly out of sample. To allow the actual forecasting efficiency to be properly evaluated,
I have also provided a series
with the actual historic series of
forecasts, which were made solely with data available for the date of the forecast. The chart below
shows the use of the Kolmogorov-Smirnov test to evaluate whether the actual factor returns are consistent with these
fully out-of-sample forecasts. For these five points, the answer is yes — as we have a 65% chance of drawing
a larger Dmax deviation statistic by chance.