dTheta/dT

A graph of theta over time, especially comparing strikes, reveals some interesting effects.
The various curves show the theta for call options when the underlying is unmoving at 1300. Notice something interesting? There's a point at which the market "gives up" on a particular option--the odds of it coming into the money just get exponentially worse, which causes theta to drop off.

This chart shows the point at which theta's Rate Of Change (or dTheta/dT) goes negative. In this case, the market "gives up" on the 1340 call (40 points out of the money) 13 days before expiration.

Trading options is almost exclusively about risk management. Looking at dTheta/dT reveals the balance point between probabilistic risk and reward.

One way of using this:

1. Sell options which are on the verge of being "given up" upon. You receive an instantaneous maximum theta, but delta and gamma are quickly dropping. Your risk is directly proportional to gamma (notice how SPAN margins penalize you for having high negative gammas, no matter how far away from the money you are). Delta quickly decaying means you'll have to make fewer position adjustments. In other words, you're finding the point at which risk is at its maximum decay and your position will be easiest to manage.
   
2. Buy back short options which no longer have enough theta to be worthwhile and roll into strikes which meet the criteria of #1. While your risk is decreasing, so is your earnings/short contract (and earnings/short contract may be the single most important parameter to optimize when short!)
   

Of course, there are arguments against this approach. The most compelling is that the Black-Scholes model (upon which theta here is computed) does not properly account for the leptokurtic movement of prices. Various authors (including Benoit Mandelbrot in The (Mis)behaviour of Markets and Nasim Taleb in Fooled by Randomness) argue that you can win in the long term by buying extremely out-of-the-money (OTM) options. You'll lose 95%+ of the time, but the one time you win, you'll win big enough to make up the difference. In the case of Nassim Taleb, he made his money buying extremely OTM options right before the 1987 stock market crash.

Personally, I'm positive of leptukurtosis in stock pricing, but that it's just another risk to manage when selling options (and a measurable, compensatable one).