A single observation of extreme value can't reveal its own noise-to-signal ratio, but repeated observations can - by how far values pull back toward the mean

Single sample doesn’t tell us how much noise-to-signal ratio does the sample carry, but we can estimate it by checking multiple samples and seeing how much closer the values are to the mean.