From his talk at Wharton:
There's only one intelligent form of investing, and that's to figure out what something is worth and try to buy it for less. Distressed debt investing is not different in that regard. We have to do the same thing. Simply put, you look at a company, you look at the business, you figure out what it could make in a normal environment, and you figure out what that company would be worth—generally to a strategic buyer once its problems are largely resolved and once the capitalization has been restructured. Then you think about how that value will be divided up among the various classes of claimants, and you figure out what a piece of a claim is worth and you see if you can buy it for less.
If you can make those judgments on the basis of conservative assumptions and still end up with good room for profit, then that's a source for margin for error. I think that the margin for error comes primarily from being able to use conservative assumptions, and then still be looking at a generous rate of return.
Now, I must say, it's not true that the more conservative the better. Because you can get to the point where you can make assumptions that are so conservative that you'll never lose money, but it will give you a target buying price that is so low that you'll never buy anything. So you have to gut it out and be willing to include some optimism, or else you may never get to buy anything.