Hi For a stock I want to create deciles of implied volatility using McMillan's formula from page 467 of the third edition of Options as a strategic investment. This will be based on 500 days of data with the "overall" IV averaged over 20 days. Essentially the implied volatilities are weighted for volume and "moneyness", with strikes 25% out disregarded. My questions are: a) I notice the VIX doesn't calculate IV in the eight days up to expiry. Is this because the IV can get too erratic and hence is disregarded? Is it good practise to forget the last week or so when calculating historical IV? b) It isn't clear in McMillan's example but I assume you calculate the IV for both puts and calls within the 25% boundary? Thanks Matt

Unfortunately I don't have the McMillan book so I can't look at his suggested formula, but I'll answer as much as I can. Yes, the whole thing breaks down close to expiration. So I would indeed forget the last week or ten days when calculating historical IV. Personally, I would calculate the puts and calls ATM, but at the other strikes I would only calculate the out-of-the-money option and ignore the ITM option. That's because the OTM is the liquid option and the one that leads. The ITM just follows. I don't have extremely strong feelings on this one but on balance, that's how I would do it. I don't care for the 25% rule. Too arbitrary. I much prefer the VIX criteria, which is to include every option with a bid greater than zero. If you DO use the 25% rule or some arbitrary cutoff, you MUST add a rule that you eliminate any strike where the OTM option has a bid of zero. Otherwise your results will be distorted.

Adding to DMO's comment. Look at McMillan's web site. He calculates a curiv (current IV) that he posts (for free) once per week. http://www.optionstrategist.com/free/analysis/data/index.html That may be all you need. He also offer daily data, for a fee. Mark http://blog.mdwoptions.com/options_for_rookies/