The results from PlanetLab et al. may not be deterministic, but you should be able to reach the same conclusions based on repeated experiments. Otherwise, your

The results from PlanetLab et al. may not be deterministic, but you should be able to reach the same conclusions based on repeated experiments. Otherwise, your results may be a little too fragile to form the basis of sweeping conclusions. This may not be ideal, but it is no worse than any branch of science which is not purely digital.