Which would be the most appropriate models to find volatility trading opportunities (i.e. plot a theoretical volatility smile I can rely on) for the following instruments:

  • Options on equity
  • Options on commodity futures

I know there may not be a clear cut answer, but any insight or guidance is appreciated.

  • $\begingroup$ With "a smile that I rely on", you mean one that you get for free for your use, or general, no matter the cost involved? if cost is not a concern, for equity, voladynamics.com/marketEquityUS_AMZN.html may be your choice. $\endgroup$
    – AKdemy
    Commented May 4, 2021 at 22:52
  • $\begingroup$ Are you familiar with curve fitting techniques, basically they take a set of market observed IV points and draw a smooth curve through them as nicely as possible. For IV, the SVI curve invented by Jim Gatheral is often used. Many references to it on this site. $\endgroup$
    – nbbo2
    Commented May 4, 2021 at 23:48
  • $\begingroup$ @AKdemy The former would be the preferred option, although I'm not scared of computing the model myself. $\endgroup$
    – Alex
    Commented May 5, 2021 at 1:28
  • $\begingroup$ @noob2 I'm familiar in the sense that I know what they are. Would this be sufficient to find trading opportunies? I was thinking about computing a model based on realized volatility and using the theoretical prices to get an edge, but for sure fitting the IV curve directly would be easier. And I can't help but think that easier means no edge... $\endgroup$
    – Alex
    Commented May 5, 2021 at 1:34

1 Answer 1


You are competing against thousands of firms, many of them doing this professionally and employing people like the ones you see in the Vola Dynamics link I provided in the comments.

So my answer is, no you will not find trading opportunities. I go even further and claim you probably never will (on your own). If you use vendors liker Bloomberg, SuperD, and whoever else you can think of, you will also not find trading opportunities based on their, arguably (fairly) sophisticated, Vol Surfaces.

Realized vol is a single number per period. How do you intend to construct a surface from this? Generally, Black Scholes assumes a lognormal distribution of returns, meaning that the logarithmic return is normally distributed. In a nutshell, the Vol Smile mainly exists because this is not true. FX actually directly quotes IVOL in a way that adjusts for skewness and kurtosis. Meaning you have different Vols for different moneyness (delta) levels you look at.

For equity, the Vola Dynamics surface is really interesting (I assume also for commodity but I do not know enough about this asset class). SVI is a frequently used tool as @nnob2 stated. There are also mixed lognormal approaches but SVI will almost certainly beat that. It is not just curve fitting though. A large junk of work will be getting the data sorted. How to compute implied dividends and forwards, how to de-Americanize options that are quoted as American, how to get a reliable interest rate curve, how to handle illiquid quotes, missing data, outliers, timing differences (some underlying assets and options on the asset trade during different times or on different exchanges).

For commodity, there are at least two additional problems; the so called Samuelson Effect and seasonality. It will all depend on what commodity you look at though as precious metals are generally always handled (quoted in IVOL) like FX. Crude oil is essentially non-seasonal. On the other hand, seasonality is very prominent in natural gas markets. You will find lots of noise and gaps.

Some food for thought can be found in Timothy Klassen's linkedin page.

  • $\begingroup$ I'm looking at Vola curves and I'm wondering, how do they find opportunities if they parametrize whatever they need to fit the implied vol perfectly? That boils down to assuming the market is 100% efficient, so I'm guessing they're not really interested in over or underestimated IV? That's what I'm personally interested in. $\endgroup$
    – Alex
    Commented May 5, 2021 at 15:06
  • $\begingroup$ I think you misunderstood what I wrote. I meant that using realized vol will be useless and tried to provide a few ideas what will be needed to find what you try to find. $\endgroup$
    – AKdemy
    Commented May 5, 2021 at 15:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.