Ok I feel a bit silly asking about this, but I have gotten myself confused and need some help sorting something out. In QFT the operators of observable quantities commute over space-like intervals, so we cannot propagate effects faster than light, causality is safe, cool. But I have somewhat convinced myself that it is perfectly ok for a single particle emitted at one place to be detected elsewhere after less than the required L/c time period. Right? Because Feynman propagators are non-zero outside the lightcone? Not to mention that the wavepacket of a photon is spread over space, which means that the flight times of a series of identical photons should follow some -probably gaussian-like- distrubution with a mean of L/c? (even in the absence of all equipment-related uncertainties) Or have I just misled myself here? Perhaps there is also some interesting interpretation of such an effect, something like you aren't really measuring the 'same' photon, it is just the there is some non-zero probability that the photon field will fluctuate in just the right way so that a photon pops out of the vacuum at just the right time to make it look like your original one went a little faster than c (or slower an equal amount of the time of course)?