Well, that too. But MAW, the main thing about which you expressed skepticism was my categoric statement that there has NEVER been a time when the SPX rose sharply over a period of days and the VIX rose too. You pointed to the extremely volatile up moves of the late nineties as times when the VIX must have gone up along with the SPX. But I haven't seen any charts showing that to be true. Not to beat a dead horse, but I've never seen the VIX fail to fall on an upmove in the SPX - regardless of how violent that upmove was. I've posted many charts demonstrating the tight inverse correlation between the SPX and the VIX. So far no one's posted a chart showing a period where that did not hold true. Nitro did manage to find a two-hour period one Friday where the negative correlation did not hold. But if that's the best anyone can do, the rule is proven.
http://www.investopedia.com/articles/optioninvestor/05/012605.asp there are plenty of sites that will show you risk graphs for options. You may want to get "think or swim demo", pretty highly rated among option folks. Remember in this volatility trade smaller size as you will see fast moves so overtrading can kill you in this market. I found working with a black sholes calculator helpful as well. This allows you to see what price an option would be at given the parameters with different inputs. http://www.numa.com/derivs/ref/calculat/option/calc-opa.htm
My skepticism was about your statement that implied volatility always decreases as the market goes up and increases as the market goes down. To support your point you brought VIX behaviour as an argument. My point was that VIX is not a global market volatility index, this way is not able to show middle and long term implied volatilities behaviour. One can't derive global implied volatility behaviour with just a short term point of view like VIX.
I see. I thought we were both talking about VIX but perhaps that was a misunderstanding on my part. I would be surprised if the inverse correlation between the SPX and the VIX did not hold for back-month IV's as well, but that one I cannot prove. It's true that there's a difference between what the VIX really is and what most people think it is.
MasterAtWork He's right. If the market thought the volatility implied in the option was too low (too cheap an option) the option would be bid up until that was no longer the case. Vice versa for what the market thought was too high implied Vol. The volatility implied in any option is the market concensus of future volatility. It can't be anything else !
Looking at Friday's settlements on Dec. ES options - the 500 puts settled at an IV of 75%. The 1100 calls settled at an IV of 44%. So if IV is the market's consensus for future volatility of the S&P 500, is the market expecting 75% volatility or 44% volatility? An option's implied volatility directly measures just one thing - the market's desire to own that particular option. Clearly, there is more demand for the 500 put than for the 1100 call. That difference cannot be explained by expectation of future volatility, since both options have the same underlying. The difference CAN be explained by the fear of a drop in the S&P, and the desire for protection against such a drop.
DMO Regarding expected future volatility I was actually refering to ATM options. But the principle of market concensus equally applies to any skew - namely that the concensus is that markets crash down much more often and severely than they crash up. From the IV's that you quote you could say the market expects the probability of <500 expiry to be xxx% and the probability of >1100 expiry to be x% and build an implied distribution for that particular market. My point, and only point, is that the market prices options (read IV) that fully reflect the expectation of future volatility, including any fat tailed distributions. Whether that expectation is correct or not in hindsight is another debate.
Well whatever works for an individual is, of course, correct for him. If looking at IV as the market's forecast volatility or implied distribution puts money in your pocket, I can't argue with that. Personally, and for whatever it's worth, that approach - while interesting and consistent with the mathematics - has never made me a penny. What I have found EXTREMELY useful though is looking at implied volatilities purely as a window into the emotional state of a market.
DMO I don't see what else you could possibly use as a volatility forecast (?). The best estimate (IMHO) is the marlet's estimate.
I'm a bad option trader and I'm forced to close a large part of my short calls for example. Because I have to do that right now, the price I'm going to pay is a pretty high offer. Of course it doesn't mean that I'm expecting a higher future volatility. How can one know that option price is not now reflecting expectation of future interest rate or dividend moves rather than volatility one ? One can't. The only thing one can see is price. The rest is model dependent and so does the interpretation. Implied volatility is model-dependent. That means two different models, for the same market price at the same time, two different implied volatilities. Volatility is model dependent. Standard deviation, variance, mean absolute deviation, average true range...are different ways to calculate different volatilities for an asset. Different numbers extracted from the same datas (prices). Implied volatility is just a common tool to erase time and underlying effects from the value an option.