|
PureBytes Links
Trading Reference Links
|
I agree that you could do a much better job without GARCH for predicting
volatility. My solution is something like this:
When Volatility is not TOO HIGH or TOO LOW, use ARIMA model or even a Linear
Regression of Time series to forecast. When Volatility is TOO HIGH or TOO
LOW, you will need to include 2nd derivative (acceleration) in the
forecasting equation. This will allow both the trend-following and
trend-fading approaches to work at the appropriate time.
Here are the problems:
-You cannot profit from correctly forecasted volatility. You can only
profit from correctly forecasted IMPLIED volatility.
-Fading Implied volatility has killed and bankrupted many traders.
Specifically, selling HIGH volatility is usually done in the process of a
short squeeze. Your mark-to-market short position will get worse and worse
as traders are forced out of their already rich positions short options. The
position will only become profitable once all the "puking" is done. It is a
feedback mechanism. If you have VERY deep pockets, then you will survive
over the long run.
-Jonathan Ludwig
|