When I identify an opportunity for a directional play, how do I decide a) what type of trade I want to do and b) what strikes do I want to do? 

Typically, I look for some sort of vertical.  This lowers my cost basis (granted it caps our potential profit).  But most importantly, I attempt to “target” a specific strike. 

I do this by “triangulating” a few approaches.  First, especially if it is a non-earnings play, we can identify support/resistance levels.  Simply have your short leg of your vertical coincide with the technical support/resistance level. 

But you can kind of throw out technical levels with an earnings report.  The fundamentals far outweigh any s/r levels. 

So, what to do?  I have two ways, one is to look at prior earnings performance.  The past is not necessarily a predictor of the future, but it is a piece of evidence, especially if the variance of the reports is low.  So, if the average move of a stock is 5% and the past six are:  4%, 5%, 6%, 4%, 5%, 6%, this is much more predictive than one like: 0%,10%, 0%, 10%, 0%, 10%.  Each has a mean value of 5%, but the distribution is much more volatile in the second example. 

Our last (and strongest) piece of evidence is what are the options telling us that is priced into the market?  Using the at-the-money implied volatility, we can calculate what move is priced in.  If the average move is 5% and the market is pricing in 10%, you cannot just ignore this.  Someone (the market) at least thinks they know something and you have to adjust your risk tolerances accordingly.