Funny when the word "absolute" and "prediction" comes into the world of trading... PM me when you find the holy grail.
You cannot predict future implied volatility, forget any double talk of GARCH and what have you . You can make an estimation of what you think IV will be but you cannot predict as that means you for sure the future IV. If you can predict future IV then you can predict prices of options. Since market makers themselves cannot predict future IV, why would anyone here claim otherwise.
I completely agree. Better to spend one's time trading IV with straddles or condors GARCH in this application is pretty much glorified technical analysis (at its worst).
I think there is some confusion here on what GARCH and IV actually are. Please consider that when you make a stupid statement about the merits/flaws of a mathematical model, you best *understand* the model (especially a Nobel prize winning model). You know who you are... Implied Vol is a "backward" solution to the Black-Scholes equation using the current option price. GARCH is a volatility forecast using best-fit parameters that maximize the log-likelihood of a data set. GARCH is a conditional variance in this sense that can model volatility clustering. Note that GARCH is a form of curve fitting in that sense as well. GARCH forecasts are actually quite good for certain markets and there exist arbitrage profit opportunities when the forecasts are different from the IV's. I believe Robert Engle, who founded the original ARCH process, won a Nobel prize for it. Mike
It cannot be done. Only thing what can be forecasted are when tops and bottoms of price occur , time when they occur. If you want to trade options, you will have to learn how to forecast tops and bottom on the underlying so you do not have to mess with condors,buterflies and other complicated strategies.
Of course predicting tops and bottoms can't be done consistently by anyone, and therefore the need for "complicated" strategies. BTW: Here is some C++ code that uses the Method of Bisection to calculate IV using Black-Scholes: http://finance-old.bi.no/~bernt/gcc_prog/recipes/recipes/node7.html#SECTION00724000000000000000
As you put it, GARCH is a forecast based on best-fit parameters but it cannot predict future volatilities. It is a model. Black Scholes also won the Nobel Prize but we all know B-S is based on unrealistic assumptions, however it is the best we have. Notice the one major flaw in B-S is its estimations or assumptions on volatility. You cannot predict future volatility, you can use models to come up with your estimates of future vols but that is as far as it goes. How many market makers are constantly running their models to make volatility estimates as well as hedge funds who trade options doing the same thing. They are all doing it because everyone has changing opinions and they trade or hedge off of those changing model outputs as factors change minute by minute in the market.
It can be done consistently, by the few of us only. best known would be AIM top two traders Max and Khan.