Well conservation of momentum and energy still apply so things don't accelerate without interatcing.

This is a fallacy. You say conservation of momentum forbids the neutrino from accelerating. This kind of deduction rests solely on classical mechanics (meaning non-quantum) and is irrelevant to Quantum Mechanics.

Your right the neutrino is described as a wavepacket delocalised over all space. Still by far and away the most likely place to find it is straight ahead.

Hence, the semi-classical approximation I was referring to is a good one.

But because it is light and waeakly interacting you get effects that classical physics won't predict, they only come out ot of wave mecahnics.

ORLY? Like what? Do you observe neutrino interference in your detectors?

No conservation of momentum doesn't forbid the neutrino from accelerating how do you get to that? The neutrino can accelerate in a manner by scattering off of other particles with weak isospin. Its true that the classical idea of an accelerating potential has to go through some levels of quantisation before you get to weak potentials but I guess on some level the principle is similar.

Semi-classical usually means that you are using some aspects of classical and quantum mechanics in a model. Typically this is where fields are classical and particles are quantised, because of the difficulties of full quantum field theory. Show me (i'm only going to accept a mathematical answer) where the classical component is in the wavepacket description of the neutrino.

Yes we can observe effects from interference of quantum properties of neutrinos.

Phys. Rev. Lett. 100, 221803 (2008) arXiv:(hep-ex)0801.4589

Shows that only neutrino oscillations can now be made to fit the observed data. Of course feel free to come up with another explanation.