Apologies if the answer to this is obvious to veterans.. as a rookie I'm not seeing it.. Suppose you sell an European OTM put on an index. Then sh*t hits the fan and 3 days before expiration the market is way down and your put is ITM. You buy back the puts at a loss but offset the loss by selling the next month's puts at the same strike price. In other words, you get a credit. Repeat every month until the puts are no longer ITM. I'm obviously missing something .. otherwise it would be possible to never get a month with a debit.
Depends on how far OTM is the put that you're selling. If you're selling it far OTM, the credit you're getting is very little compared to what it's going to be worth when it gets ITM. For e.g. assuming IV stays constant throughout the trade, and you sell an 8 delta put on RUT today, for $2.35, that same put will be worth $19.55 when it gets in the money. That's a very unrealistic price even, because if the underlying were to move down to hit your 8 delta put, the IVs are surely going to get jacked up and your 8 delta put may be worth $23 - $24 if IVs were to go up 5 points. It may be worth even more. Then there's no way you're going to roll this out the next month at the same deltas. You're going to have to come much closer in, to a 33 - 38 delta put. This also means your put will have a much higher chance of getting ITM. The probabilities of the underlying touching these newly rolled puts before expiry, are 73.76% and 82.77% respectively. If this gets hit again, it's simply going to be a Martingale kind of strategy where you'll have to keep increasing contract size in order not to go ITM. If you play this game enough times, it's a matter of time before blowing up.
But what happens if you keep the contract size the same and sell the exact same strike price for the next month, even if it is ITM? Assuming the worst case, let's say the put is far ITM, and its value is now $30. The next month's put at the same strike price is $30.10 because it's so far ITM. So by rolling you just avoid paying the piper until the RUT gets back get where it was. You will have a very poor (but still non-zero) gain during the time you are ITM.
Firstly, your margin requirements to sell that same strike put ITM in a further out month is going to more than double from what it was when you first sold it OTM. If the underlying really tanks, then the margin requirements may go up even more than that. How much margins are you going to set aside when you first start trading this? I think this is one major problem. If you get a margin call and you can't put up, your position will be liquidated. By rolling the put, you will lose money from slippage, and commissions. The tiny time premium you make will not even cover that. And because the put is deep ITM, the slipage is going to be much greater than when it was far OTM, because now the bid-ask is going to be of a much higher value. For e.g. when it was far OTM, the bid-ask may be $0.40 - $0.60 (20 cent spread). When it's deep ITM, the bid-ask could be $91.40 - $93.40 ($2 spread). By the time it takes for that put to get OTM again, you'd have lost a lot of money from slippage, and some from commissions. In a worst case, the put could get so far ITM that it's no longer listed for trading. Unlikely, but a possibility. More importantly, is the ROI of this strategy worth trading? If the puts ever get ITM, you're tying up a lot of capital which is not working for you, and there is no telling how long it will be tied up there. And as the underlying tanks more, you're going to need more and more capital to be tied up not working for you. Psychologically it's kinda hard to do that, throwing money after a bad trade. But other than these points, I could not think of any other way this could not work. But it's an interesting idea.
What if you were to do this with a vertical (bull put spread) instead? That would keep your margin requirements the same throughout the entire process, so you wouldn't alter your position size. Your points about slippage / opportunity costs are well taken, but wouldn't those be a small price to pay to avoid the pain of buying back .90 deltas when you sold .08s? The intent of the strategy is just to dodge a large loss while you wait for recovery. Worst case you end up with a 5 year bear market .. and your money does nothing during that time.. but at least you don't lose.
European style puts that are deep ITM trade below parity because they can't be exercised until expiry. You won't be able to roll for a credit. Even American style deep ITM options often cannot be sold for parity because they aren't liquid. In reality, rolling for a credit or even BE may not be possible. Aside from that it's not a good idea because of the ample reasons already provided. This might be a viable strategy for you, but get some experience before trying it. The spread only complicates matters and you may end up paying too much to roll. Such is the reality of the markets.
Margins for verticals will also change. The margin req'ments are the strike difference between shorts and longs, minus the credit you recieve. Just a suggestion, but you may want to re-evaluate the mental bias of trading a strategy that "cannot lose." It's an attractive proposition and one I'd admit I'm very prone to also, but I don't think it's productive (unless you're doing arbitrage). Say you trade this way, and on average you make 5% a year on your capital over the long term. (just a wild guess) Versus a strategy that takes losses 30 trades out of 100, but makes 60% ROI each year. Which would you rather trade? With the second method, you take your losses, then use the remaining capital to trade again the next month. Versus the first one where your locked up capital does nothing for you except lose money and you have no idea when it's going to stop losing. Slippage is going to be worse when trading a vertical spread, vs. trading a single option. I guess you need to figure out the expectancy of your system first, and determine the position sizing method you're going to use, then decide if it's worth doing. I think the only way to get an indication of how your system would perform in real life, is to backtest it for 3-4 years, but pay particular attention to modelling with reasonable slippage. It may be profitable, or it might not. But it's worth doing so at least even if it didn't work, you'd have learned yet another way that doesn't work, and possibly learned many other things in the process.