I think you're missing my point. We can conclude that the accused is guilty, by your measures, just in case the accused is actually guilty, since you want to rule out faked evidence. But that makes your evidentiary standard meaningless, since it's just a complicated way of saying "that guy is actually guilty".

Now, if you want to talk about whether the death penalty is permissible given an uncontroversially guilty agent, that's a different kettle of fish entirely. You might find Montague's Punishment as Societal Defense enlightening here, or any of the proportionality literature by Jeff MacMahon.