This gave me the biggest headache and there's no easy answer. A simple strategy was vertical spreads for me.
a good friend of mine that used to work on the floor of cbot for many years says "buy the meat sell the wings"
The more your options are at-the-money, the higher their IV, and the less time remaining, the higher your theta The more your options are out-of-the-money, the lower their IV, and the more time remaining, the lower your theta.
Well now, that's a mighty strange way to reduce theta burn. Mighty disadvantageous way to position yourself for lotsa reasons.
in "options made easy", Guy Cohen advises you close all (long) options with <=1 month left, as theta decay accelerates rapidly with 1 month left. personally, for short term long options, I buy them ITM with ~ 3 months time left.
I agree with DMO, that's weird. All of the convexity [delta and vega] is in the wings. Most locals are natural buyers of flies, not sellers.
Personally (not that I would claim to be any sort of expert), it's not like you're paying theta for nothing, is it? You generally get the gamma you pay for, as otherwise things would really be too simple. So I don't think there's a hard and fast rule. In my view, the only way to decide when you're paying too much theta is to actually perform the analysis. If, given the parameters of your situation (hedging freq, transaction costs), you find out you're paying too much for the gamma with <1mo to go, by all means get out. My Z$2c..
Actually in equity options the out of the money puts will trade with a higher IV and taking that one further the winger calls or puts will trade at a higher IV then the AT's