Re: trying to find fat tails

Naveen,
I was trying to find "fat tails". I looked at NG prompt month prices'
log-returns for 5 years 9 months.
On the figure below you can see the comparison of empirical cumulative
probability function with normal
cumulative for this time series (standardized: mean subtracted, divided by
stdev). The effect of fat tails
is not pronounced so much. The "fat tails" effect was much more visible on
your plot when you looked at the
oct-00 prices log-returns versus my time series of prompt month's prices.
The shape of the distribution is different from normal, though, and fits well
with the volatility switching model.

Tanya.