Trading a select group of options is something that I am constantly thinking about. Currently, I do not have a select group of options. Instead, I download information each morning into an access database that I wrote so I can seek out very specific details on options.
One thing I look at almost all the time is the percentile of implied volatility AND the gaps between the historic and the implied, both in composite and skews. This is what I use to determine what are "cheap" and "expensive" options. When I find an option with implied volatility below the 5th percentile AND with historic vol trending up and currently above implied volatility, I feel safe that I'm looking at a genuinely cheap option.
This is basically a modification to Larry McMillian's approuch. He tells traders to trade ALL markets (futures, equities, indexes, etc.) And look for extremes in volatility. The story goes that for the most part, implied volatility maintains a range and it's MUCH easier to identify a true high and low. In theory, it sounds great.
Now here's the problem with this approuch. Though I'm only looking at trades with high probabilities of success, they are typically not liquid. The Bid/Ask spreads on these trades are usually rediculously wide, so what ever benefit you get from the finding an option at an extreme is usually lost to the added risk of dealing with a wide bid/ask spread. So, you wind up waiting for an extreme to occur in something that has enough liquidity to trade. In the end, you come up with very, very few trades - too few in my opinion. The only exception to this is perhaps long straddles.
Mr. McMillian does not discuss this MAJOR drawback in any of his books and it's my biggest problem with his method. I can go for days without finding a single trade and I'm using a computer to find them. More and more I find myself considering the benefits of building a "core" group.
Has anyone else been down this road?