I wanted to get some input on my design for measuring and forecasting volatility, I start off looking at the current 10 day, 20 day, and 30 day HV taking note of the trend, wether the volatilty is increasing or decreasing. I then forecast the volatilty up until the expiration of the contract I am looking at, I calculate it using GARCH. Based on this analysis I eliminate trading strategies that don't apply to the forecast. ie the forecast calls for a decrease in volatility, then I sell volatility etc. Anyone care to share their logic and methods? I am having a hard time coming up with a system, seems like I am reinventing the wheel from bits and pieces.
The best way to forcast volatility is simple historical volatility. Anything else is just distracting.
This is not true. Suppose that you were playing a biotech stock that is trading all over the place and the historical volatility is very high. Now suppose that you and the rest of the world knows that there will be a FDA meeting on a certain date. If the FDA approves the company will be worth billions. If the FDA does not approve the company will go bankrupt. Now if you are using straight historical volatility, you would sell the straddle and expect to make a lot of money. However, because you haven't taken into account the FDA meeting, you will instead be taken to the cleaners.
freehouse-you talking about totally different things. you talking about IV based on event, and first dude talking about forecasting IV based on HV and ask, if he(she) on right track)))) Freehouse, i know, what you are talking about-VPHM)))))). if you sell straddle on that day( before FDA)- you not going to lose much. bw-the news about the meeting was out a month before it. and if you bought it on that time for a $1 on both sides, you will sell it at 4 on each side on day before FDA. GILD have FDA meeting on 08/06. try to play there)))))) IV is very tricky and i would say, it is impossible to predict it. you figure out IV of the stock, - you say it will go down, then market start move down and VIX start move up/ what then you going to do? HV-just calculation of past price movement, did not give you much. most of time IV depends on event or IV of the market. ---Riesgo2002! how you predictions so far? and what is GARCH means? Good luck all!
I'm not looking to predict anything, I just want to be able to guage option IVs against the HVs of the underlying, at the same time I want to have an idea of where the volatilities are going in the short term. GARCH is a mathematical model used to forecast volatilities based on HVs, it is an autoregressive model, meaning that it reverts back to the mean over time. Anyways I know very little of the nuts and bolts of how it works, or even if it will work for my purposes hence this thread.
GARCH is a waste of time. All it does is scramble volatility on itself and time. It forecasts VERY bad. There are very few who are capable of 'leptokurtic, skewed distributions with multiple blah blah blah' that is actually useful. GARCH, STARCH ARCH and all that crap is a waste of time. Black Scholes is worse. For all PRACTICAL purposes, for the VAST majority, average historical volatility and a little common sense is the best way to estimate volatility.
I look at a lot longer term than 30 days. I look at IV as well- for example, to buy I'd look for an underlying with options trading in the lowest 20% of its 6 month IV and SV range. (Obviously, I'd check the news for a reason for its low volatility- such as with TRW- which unfortunately is the only stock that would have otherwise qualified today)
Riesgo, I did a study several years ago on predicting Volitility. I used percents of historic vo, imp vo and the vo over the last 10 days and compared them to the vo for the next 10 days. The best combination was 40% imp, 40% 10 day vo and 20% historic. Hope this helps. jj