Apologies if this is a repost but I did not see it discussed here.Does anyone have any recommendations for calculating implied volatilities from a strip of quoted options? The strip is on the same underlying but with different maturities. There is certainly no unique solution and I was thinking about trying to solve this optimized to minimizing the change of the IV surface from the previous days settles. I am concerned that this could potentially dampen the effect of actual market dynamics attributed from a single contract contained in the strip but at the moment I am not certain how else to include this information besides from an ad-hoc trial and error.Any thoughts would be appreciated. Thanks!!

You might try to reduce the strip information to a "standardized" implied volatility.See the CBOE's VIX (Old methodology) white papers.

Alan thanks for the response ... I might not have given enough information, sorry about that.The quote for the strip of options is the average price of a strip of 3 - 12 months (one option for each month). The VXO (old VIX) assumes the term structure of volatility is linear and interpolates between the observed volatilities.Since I am observing the average option value, I cannot use an estimate of a standardized implied volatility because of the non-linear payoff of each option. I instead have to adjust the IV of each option to ensure that I am pricing withing the spread. In doing this, I could create an artificial (or inaccurate) term structure and still be priced within this spread, this is what I am trying to avoid.I can build a skew from daily settles, restricted to the bounds as outlined in Lee (2002), however, I would like to update the skew from these observed quotes of strips but am not certain how to do this in a meaningful way other than adjust on aggregate. This may be all I can do, just seems intuitive that most of the adjustment should typically be in the front end of the curve and not even over all terms.

I think I see. You observe a single daily last price (settle) for a strip. Apparently, you have market quotes for eachcomponent of the strip also, and you are simply trying to infer prices/implied vols for the components.Is that right? What I don't get is this. Is the strip price within the market bid-ask of the components (i.e.adding up all the market bids of the components, and similarly for the sum of the market asks)? If so, I don't see the issue.

I have the settles for each component. I observe market quotes for various strips. Typically my marks are withing these quotes but as the market evolves I observe strips that my marks don't match and need to update my surface. If I observe a 12 month strip I have 12 vols that I can adjust to satisfy a single pricing constraint. I'm looking for ideas on how to update my surface in the most meaningful way without arbitrage opportunities.Currently the only ideas I have:1) Update that surface to fit within the spread subject to minimizing the difference from the original shape. (This would most likely have an effect over the entire term structure which is not necessarily intuitive to me)2) Use some sort of weighting for the front end of the curve. (Fairly Arbitrary and can greatly effect prompt skews)3) Go for lunch and pretend I didn't see the quote. (There is a great Thia place nearby!) All three seem fairly equivalent to me, hence, I'm not really satisfied with my current ideas and I am looking for something new.Thanks!!

Well, I don't know your market, so am just throwing out ideas. Based on what you have said, it sounds likeyou know how to update your vols once you know the next settlement prices for the components.So, it is a matter of predicting those next settlement prices, given the intra-day move in the strips quote.In turn, that sounds like a regression problem. Use all the historical data that is available (probably keepingsome for out-of-sample tests) to create a good predictor of the next component settlement prices. Then,update your vols as if those predictions were accurate.

Hi Alan,Yes this has been my problem, and I'll give your suggestion a try. I appreciate all your input.Cheers!