Forecast volatility
- Thread starter Gustaf
- Start date
Quote from Gustaf:
I would like to have some volatility predictor for my option strategy evaluator, anyone have any idea how to predict.
It can be simple or complex, psuedo code is fine too.
br
When you say "volatility" are you referring to the volatility implied by option prices or the volatility of the underlying asset? The common "naive" forecast is the most recently observed volatility held constant over the out of sample forecast period.
You are definitely correct that forecasting future volatility is the key to modeling option positions.
While it may be true that statistical analysis is the only way to measure true recent volatility, market implied volatility (IV) is generally considered a better indication of future volatility than the statistical method, because it is sensitive to current market perceptions.
Options are currently priced to reflect the combined expectations of all market particiapants about the future volatility of the underlying asset. These expectations are evident almost instantly when you measure volatility the "implied" way.
So that is what most people use. When I do something like look for a strike one standard deviation away from the current market price, I use the IV of the ATM call option in the month I am analyzing to create the lognormal distribution.
Implied volatility is, admittedly, somewhat circular. If everyone relied exclusively on implied volatility, the slightest error (i.e. rounding errors) could cause everyone to drift further and further from the true volatility of an underlying asset until there was no resemblance between assumed and actual volatilities. So it never hurts to look at statistical volatility to make sure you are "staying on course".
While it may be true that statistical analysis is the only way to measure true recent volatility, market implied volatility (IV) is generally considered a better indication of future volatility than the statistical method, because it is sensitive to current market perceptions.
Options are currently priced to reflect the combined expectations of all market particiapants about the future volatility of the underlying asset. These expectations are evident almost instantly when you measure volatility the "implied" way.
So that is what most people use. When I do something like look for a strike one standard deviation away from the current market price, I use the IV of the ATM call option in the month I am analyzing to create the lognormal distribution.
Implied volatility is, admittedly, somewhat circular. If everyone relied exclusively on implied volatility, the slightest error (i.e. rounding errors) could cause everyone to drift further and further from the true volatility of an underlying asset until there was no resemblance between assumed and actual volatilities. So it never hurts to look at statistical volatility to make sure you are "staying on course".
Uh oh... Someone says "vol forecasting" and Segv is all ears!! LOL
BTW, to the OP, Segv is probably the best person to help you out on this if he's willing.
Quote from Cache Landing:
Uh oh... Someone says "vol forecasting" and Segv is all ears!! LOL
BTW, to the OP, Segv is probably the best person to help you out on this if he's willing.
Iam speaking about underlying instrument volatility (Indexes in my case).
Well anyone that helps out can use my Iron Condor trader for IB Api (c++).
Iam gonna read all answers once more to contemplate. Thx guys!
This is the application Iam working on:
http://www.elitetrader.com/vb/showthread.php?s=&threadid=94354
http://www.elitetrader.com/vb/showthread.php?s=&threadid=94354
That is not quite accurate. Most of the research shows that using implied volatility produces biased forecasts of future realized volatility. More recent papers have shown that implied volatility is an unbiased estimator of realized volatility when adjusted for a negative risk premium. The "combined expectation" of market participants derived from option implied volatility is often higher than the future realized volatility.
I have spent the last 10 years working on that problem, so my current best answer is not something that I am likely to share. I will however point you towards the work of Torben G. Andersen, Tim Bollerslev, Francis Diebold, Paul Labys, which has spawned a lot of productive research over the last 6 years. Ser-Huang Poon's book, "Forecasting Volatility in Financial Markets: A Practical Guide", provides a good introduction to the research. I also would recommend Steven Taylor's excellent book "Asset Price Dynamics, Volatility, and Prediction".