Now that the volatility risk of ultrashort funds is clear to me (thanks again, MasterPlan), I've cooked up a risk-control strategy. It's a bit labor-intensive, but I'd like feedback on the possible benefits vs. anything I might have miscalculated.

I thought about just using unleveraged inverse funds, but they're not even available for many indexes, and I don't want to commit the extra cash (actually, margin borrowing, at this point).  Then I thought about just reducing my exposure, but that leaves me with little hedge against a market crash.

So instead, I'm going to monitor the leveraged funds closely, and collar them with limit orders above and below. So, if the price starts to drop by more than 1-2%, the stop order kicks in and sells off 1/2 of my holding. If it's a trend reversal, I've saved 1/2 my loss; if it's a temporary correction, I've just reduced my exposure, and can buy back in when I'm sure of the trend.  Likewise, if the index goes down more than 1-2% in one day, sending the inverse up 3-4%, I have a limit sell to skim the profits; then if the index bounces, I don't lose anything to the volatility. And if the index doesn't bounce, I've reduced my exposure proportionately (plan B).

Example: RRZ (-2x Russell 2000) closed Thursday at 94.70. I have 160 shares, with a sell order for 80 shares at a stop of 93, and another sell order for 40 shares at a limit of 99.  I expect neither order to be triggered; let's say Monday's close is 97. Then I will move the collar up accordingly, to 95/101.  I have no problem doing the daily maintenance required; if it looks like I'm going to be gone for a few days, I'll just leave a wider collar.

This looks to me like minimal risk with only modest sacrifice of possible upside-- but I have neither the experience to judge whether it's likely to work as planned, nor the mathematical tools to optimize it. Any kind of critique would be welcome.

--Aalan