10 to 1000 atoms per cubic meter is really deep vacuum and it means we could get interstellar travel by going really fast, continuously accelerating with a cyclotron drive to close to speed of light, say 80%, then turn to decelerating halfway there.
Your arguments are sound. A lot of stuff out there would mean a blue sky everywhere, as gases do Rayleigh scattering of light. And once you go past our atmosphere, the rest of the Universe is black, not blue, the blue sky disappears, therefore there is deep vacuum everywhere, or at least small particle free space, and then you have to invent something else that is particle free but present in vacuum and retards photons compared to neutrinos, such as gravity, or dark energy, or whatnot.
Photons interact a lot with "subspace" vacuum and get delayed, but neutrinos do not, or not to the same extent. If you're absolutely certain that neutrinos can oscillate, and there are different kinds that can turn into each other, then having two pulses does not make sense if they were both from the same event, as any different neutrinos should have oscillated into each other and be indistinguishable. One question, when we talk of photons, it's hard to talk about polarization, but we know light is a transverse wave and polarized. Obviously neutrinos have an associated wave-particle duality to them, just like everything else, and do we know what kind of waves they are? Longitudinal, or transverse? If they are transverse waves, then there could be neutrino polarization, and just like with light birefringence in a calcite crystal, where an incoming uniformly random polarized wave splits into a fast and a slow beam, based on polarization, so if the space between us and the supernova is anisotropic in any sense, such as gravity pointing in certain direction throughout, and neutrinos be polarized transverse waves, then there could be a fast and a slow wave with them, but not so with the light wave, unless they had polarized light receptors and have info on the polarization makeup of the light received vs. time (and this would be low intensity for a while, then intensity doubling when the slow beam arrives too and adds to the fast one, as light emission was continuous with a slow decay, but the neutrino came in pulses.) So if gravity affects the speed of light, and affects it in a birefringent way, it may also affect the speed of neutrinos, if they are transverse polarized, in a birefringent way too, and then none of the signals really arrived at the speed of light, but slower, as in a calcite crystal even the fast wave still has a reduced speed from true speed of light.
By the way I still don't comprehend the concept of how a uniformly polarized beam decides to split into two in a calcite crystal, instead of a spread spectrum, like how does a wave just below 44.9 degrees decide to go with one beam, then one at 45.1 degrees with the other beam, or is that the cutoff point, the math must be really complicated, but a lot of XIX century mathematicians well versed in such things would probably have no problem explaining why.
Also, looking at the double slit experiments, an electron is a wave that passes through both holes, then it decides to collapse at some point on the screen, how does a wave decide to become a particle, or even if not a particle, an interaction, in effect we have no particles, just waves, and they interact at given points, sometimes within very strict limits on location, such as a particle trace in a cloud chamber, sometimes in very random locations, such as where an electron collapses on a screen after having passed through a double slit. Many double slits in series of course would confine the electron to a linear path too, just like a cloud chamber cloud does, if the electron found a way to not interact with the walls in series, a sort of filtering effect. One that interacts with the wall off angle through diffraction then changes its mind and returns to being an electron on the original straight path, would be like the particle going through the cloud chamber, leaving a trail along its path, as only the electric charged ones leave a trail, but electric field interactions can be thought of as Feyman interaction particle exchanges, not as a continuum, but interaction particle abundance with a certain probability function description, and so can other "field interactions" like gravity through gravitons. There is different ways of looking at the same problem, and, like descriptions of the same phenomenon of heat transfer through flow of caloric and kinetic molecular theory both work until we find a fault with one of the descriptions, such as caloric can be generated by mechanical motion, and is not a conserved quantity, so it has no use as a concept compared to mechanical motion being heat. In the wave-particle duality there might be a similar debacle where one concept does not fit with reality, and things such as particles might have to be abandoned, and everything be thought of as a wave that decides to collapse and interact and whatever point it decides to do so, and the best description we have of where it decides to do so is a quantum statistical one - but there might be a way to measure and predict why and where such a things should and would happen, even if not practically - but at least in concept. As in practically would require measuring on the order of 10^23 atoms, and even so it's not the atoms doing it, but something in subspace vacuum that fluctuates, We don't really understand the structure of vacuum.
I can barely keep my eyes open again. Still haven't read up on neutrino detectors