If it goes ITM, then you hedge with the underlying. Net effect is about PnL=0. So what?
I bind the cash needed for hedging only a few days, as these are weeklys or bi-weeklys.
And when hedging activates then the position counts as "covered"...
I understand what you're saying, you buy the underlying if the price gets close to the strike price, which will eliminate the possibility of a "blow up", remove assignment risk, and eliminate a margin call -- provided you have enough in your account to actually buy the underlying. There's 2 key problems I see with buying the underlying when the price nears the strike:
1) You actually have to have the capital in your account to buy the underlying. This means that you must limit the number of contracts you short sell such that you can afford to buy the underlying if its price comes close to the strike (this means you need to calculate based on the assumption that the underlying's price is close to the strike, not its current price, which is much below the strikes you're shorting).
2a) The case of the underlying coming near your strike and then dropping. The options trades you've been talking about are for premiums on the order of $0.2/share. If you buy the underlying when it reaches the strike, that's fine if the price continues to go up, but what if it doesn't? You lose if the price drops below your purchase price - the options premium - commissions.
2b) You could say if the underlying reaches the strike and then drops, you'll sell, but what if it begins oscillating at or near the strike. How do you deal with this situation?
I do think that buying the underlying if it gets near your strike is a way to reduce some risks, it introduces another. It also requires you to limit the how much of your account is used on a single short option position such that you can afford the underlying. This will affect your projected PnL.
Just some things to consider. I'm no expert, but I've been doing a lot of my own analysis of options planning on automating my own ideas and I'm pretty sure I know your strategy. It's pretty simple. When you have a simple strategy you need to think, "why aren't other people doing this?" The answer is almost always because the probability of loss * value of loss > probability of success * value of success, or using better statistical terms, the expected value over a long time frame is <= ~0.