As I've pointed out many, many times, options volatility remains wildly overpriced compared to realized volatility -- whether you look at individual stocks or indices.

When there's out-of-line pricing like this, does it have any predictive ability? The great Lawrence McMillan recently took a look.

Remember, the VIX looks forward 30 calendar days, while historical volatility (HV) looks backward in trading days. So to come reasonably close to comparing apples to apples, he uses 20-day HV. Over the course of time, the VIX overprices 20-day HV in SPX by an average of 4 points.

Going back to 1993, McMillan found 35 instances where the VIX versus the 20-day HV differential was 10 points or more. Obviously, it has to resolve in some fashion -- either with HV picking up or with implied volatility (IV) falling, since the market won't overbid for options forever. McMillan found that in 31 of the 35 instances, it resolved with HV perking up.

So fast-forward to today: It sounds like odds favor an increase in stock volatility.

But before you jump to conclusions that volatility equals a down market, consider this: When those 35 instances above resolved, there was little predictive ability as far as the market was concerned. McMillan found the market lower 10 times, unchanged 10 times, and higher 15 times.

So he went another step and related HV to VIX futures and to VIX futures premiums. He found that, when you got VIX premiums to HV high as above, and VIX blended futures at a premium, it provided a good market signal. And that signal was down.

The problem is that there are only 5 instances, because VIX futures are only 3 years old.