I’m guessing far fewer resources than expended keeping the person in the system.

But that’s just the issue here – when should a court be willing to take another look at the evidence in a habeas petition? Just the act of taking another look at the evidence requires a substantial investment of resources. I’m increasingly convinced that the 9th Circuit erred in making its decision retroactive (although I’m not totally sold), but I don’t see much of an alternative to what they did in terms of proactively setting forth the circumstances under which it will be willing to take another look at the evidence.

Regarding retroactivity, though, an important thing to keep in mind is that we’re not talking about after-acquired evidence of innocence but instead evidence that the prisoner conceded was available all along. We’re also talking about a determination of innocence well over a decade after the original trial, without (AFAIK) video thereof, and where the purportedly exculpatory evidence merely bolsters evidence that was presented at trial, rather than something like DNA evidence. That leaves an awful lot of inferences that have to be made – more than a decade after the original trial, no less – to reach the conclusion that, had the additional evidence been introduced, it is more likely than not that no reasonable juror would have voted to convict (being the standard for a finding of “actual innocence”). As I say above, I’m largely persuaded that the 9th Circuit shouldn’t have applied its decision retroactively against this prisoner; however, I also don’t think it’s wise to fall into the trap of treating the district court’s determination on the merits as unassailable proof of innocence.

I understand why there’s an administrative need for finality, and I understand that convicted criminals would file frivolous appeals that would burden the court system, and I understand the need to follow rules and procedures…but, man, we’re talking about someone who seems to be innocent here. It’s got to be wrong to refuse to even look at the evidence, doesn’t it?