Is there any relationship between Average True Range and Implied Volatilities? In other words, given an ATR for a stock, should you roughly be able to calculate the IV for the options?
runningman, I don't think so. Look at it this way: ATR is recent history and Implied Volatility expresses how option players expect the stock to behave in the future. Take a typical if somewhat amplified example: A fairly active stock quieted down, there is a standoff because earnings are expected. While the ATR is narrowing implied (and to a large extent for the front month) is actually going sky high because option writers demand more premium to compensate for the extra risk for the period. Looking at Historic Volatility at the time of similar occurences in the past may show you the extent of volatility expansion at those times but no two moments in market history are exactly alike. I seem to remember that IVolatility .com has historical data on IV. Now that could be the closest to having useful information. I am curious what other posters more knowledgeable then moi have to say on this. Regards, GC
As the OP mentioned, there's no direct relationship between ATR and IV, as one is backward-looking while the other is forward-looking. So you cannot calculate one from the other. You can use both to estimate the relative cheapness/expensiveness of the options, but that would be your own subjective assessment.
As above, one looks forward, the other looks backward. But if you're going to look backwards, I think it's helpful to use the same annualized volatility measurement used in the option pricing models. If you have metastock or amibroker, it can be plotted with the following formula which gives an annualized figure based on the last 30 data points: (StDev(log(C/Ref(C,-1)),30) * sqrt(256))*100 You can also calculate it in excel. Or just use CBOE/ivolatility vol charts which plot both IV and the above HV formula together.
Thanks for the replies. For some reason I thought these numbers should be pretty much the same. I figure since IV/sqrt(256) *stock price= about the average volatility per day 2/3 of the time, that number should be pretty close to the ATR. Just a random thought I had driving around one night.
If memory serves me, Wilder compares today's High and Low to yesterday's close and also calculates today's H-L difference. He then takes an EMA of the highest of the 3 measurements. As mentioned by OP's, IV is a forward looking expectation of stock price behavior. FWIW, I think that this is apples and oranges. I don't think that there's any reason to expect that the difference b/t H and L or L and C(-1) or H and C(-1) will correlate with IV. Often, news is expected, IV inflates, yet the stock flatlines before the release. And other times, the stock gets choppy intraday but IV doesn't budge. A more reasonable connection might be ATR and HV but even that would be suspect since HV is close to close based.
HV must certainly have *some correlation to IV. It won't tell us the realized volatility in advance, nothing can, but it certainly gives us a clue as to the range of expectations of what might be realized. The caveat of course is news events. For instance (to take a ludicrous example to illustrate the point) if I were looking at QQQQ options with IVs of ~150% and HVs were running out at 10%-20% over the last 3 years, then unless there was a Cuban missile crisis MkII or something, I could reasonably determine that those option prices are ludicrously high, based on HV I would swallow razor blades than buy those options and would be falling over myself to get short on gamma. (I would cover myself just in case the market knew something I didn't though :eek: ) You can even use HV to get a clue as to what sort of vol is possible coming off news. Sometimes* those higher pre news IVs look cheap. $0.02
Absolutely true. That's the entire basis for postulating the risk/reward of an option strategy post news. But what does that have to do with IV versus ATR?
It seemed you where questioning the correlation between measured historic volatility and IV. If not, my apologies.
No problem at all What I was alluding to was that one might try to make a connection between ATR and HV. That might be more reasonable since both are historical (as compared to the forward looking aspect of IV). But even that connection would be suspect since ATR involves differences b/t the H, L & C whereas HV is based on C to C changes. $ .02125