Quote from cnms2:
If IV is in a high percentile of its past values there is no guarantee when and if it will drop.
If it follows a pattern, i.e. it is before the earnings report, the IV raised as it did before past earnings reports, and in the past it dropped immediately after the report was issued, still there is no guarantee if and when it will drop again.
You can play it, but better make sure you use stop losses and good money management.
I'm not saying that "a portfolio of straddle trades based on a screen for high relative IV would tend to lose money", I'm saying that if applied mechanically or randomly, in long run it should have zero expectancy (ignoring slippage and commissions), in my opinion.
IV is determined by the option's price, while HV is determined by the underlying's price action. The option's price is determined by the underlying's price (among others), but IV and HV don't correlate beyond chance.
Some might disagree, i.e. McMillan in his Bible.
Interesting,
It seems to me that this is what is being experimented with on one of RiskArb's threads here. I'll have to go back and read through it to see if I am missing some other criteria for the trades.