If you’re making a Pareto Optimization, it’s reasonable to think that if the probability of being murdered by the state is still much lower than the probability of being murdered by an incompletely reformed criminal for each subset of the population, there’s an argument that every subset of the population was made better off.
I’d hate to overlook a 10x improvement for all because one group got a 2/10ths penalty and another got a 1/10th penalty. The alternative (to give one discriminated against group a penalty 50x worse and the lesser discriminated group a penalty 100x worse “because fairness demands it”) is worse, IMO.
I am a little worried that you are using these made up numbers actively to justify your thinking to yourself. I understand it’s just a hypothetical to explain your framework for reasoning. But you should have actual numbers before you commit to that framing, no?
The subtext of my comment was that your formulation might be immoral given the racial context (i.e. “racist”), and I think your detailed answer still probably is that, by assigning different wrongful death targets to different races. But I understand your moral position is that you just want the best possible for each race, which is.... well, racist but morally defensible I suppose.
I believe you introduced the “what if the state murder rate was racially biased?”
I used a constant non-state rate in my example. If I had actual data, I’d have been happy to use it. Without that, I had to go theoretical and argue that if every group is made better off, there’s an argument to support. Surely, before such a plan is changed, it will be studied to an extent that’s impractical for news.yc commenting.
What if the rate of false positives varies by race, is that still acceptable to you?