It's an improvement, but not much of one. If the best you can figure out in hindsight is a way to only lose $2 or $3 billion dollars, you're not trying hard enough.