I think Holly Smith's 'Culpable Ignorance' also mistakes what culpability itself consists in. She claims (p.568) that believers in moral luck hold that an agent "is more culpable -- a worse person -- for [the foreseeable risk's] occurrence."
But that's not right: the recklessly drunk driver who narrowly escapes an accident is obviously just as bad a person as his intrinsic duplicate who (due to slightly different external circumstances) kills a pedestrian. It's just that, through a stroke of luck, he's not responsible for any deaths. This means there's less we can blame him for.* It doesn't mean he's a better person.
So the upshot of moral luck is that judgments of culpability and of character may diverge. Two equally bad people may differ in what they've done (or caused), and hence what we can blame or hold them responsible for. To be more culpable than another is hence not the same thing as to be "a worse person".
(Smith actually makes the point earlier in her paper that we blame people for their actions, not their character, so it's odd to see her make this mistake later on.)
* [I think Nate first pointed this out to me.]