I picked up a book with this very title at a thrift store this week. It's the one by Jeff Augen. Seems this person has a few credentials, but I am half way through the book and a little weary that this may be more about data-cherry picking from history, and not any real use in the present. Someone who remembers reading this prove me wrong...
Having tried this a few years back, it is intriguing but unsustainable as a core strategy in terms of having a consistent edge. The problem with very short-dated options is balancing theta and gamma. You can buy gamma cheaply at expiration but you need an extreme move in the underlying to offset rapidly accelerating time decay. Your timing and entry has to be perfect.
I've read 1 of his books about a decade ago, he seems to lack experience from what I remember. not much of the content was applicable for real trades.
Gotta be careful with the concept of expected value. If I told you that there was a contest where 999 out of 1000 draws would get you $1,001.01 and one would lose you $1,000,000 that would have a positive expected value, but most people wouldn't enter the contest.
But that is not PEV, that is a guaranteed NEV, no? If you are drawing 1,000 times, where 999 times you win the 1000.01, and one time you lose 1,000,000, then you KNOW you will be losing 990 each go around of 1000 draws. Why WOULD anyone take that?
Should have been more specific that you had a 99.9% chance of winning $1001.01 and .001% chance of losing $1M. Do that a couple million times and you'll almost certainly end up a couple dollars ahead. But the fact it's a positive expected value doesn't make it something worth risking over say, 20 tries. Although I guess people gamble at casinos with a negative expected value so maybe that's just me.
In this case, your best bet is to first look at empirical data of that particular index/ equity to estimate the probability of success/ failure. Next extend it to other indexes or equities of same nature. Also, go beyond empirical data and estimate parameters and simulate fat tail events. Last but not the least, trade small for such negative skew events. For the example that you mentioned with a tiny edge, i will probably give it a miss if this is presented in my analysis/ study. But who knows, it could be a peso event.
I finally finished the book now. I'm going to put it aside and give it another re-read over in about a year's time to see if I get a better understanding. Interesting side-note, the author actually included his drafted visual-basic-script code for anyone else wishing to duplicate and play around with crossing-strikes for backtesting.