One approach: you look at what the physical laws appear to be, and then ask "What weird edge cases would we expect to see if and only if this were running in a

One approach: you look at what the physical laws appear to be, and then ask "What weird edge cases would we expect to see if and only if this were running in a particular type of simulation?" That's what the paper linked to does. It's not perfect, but it's something.