[simonh]

Not every galaxy does follow the behaviour that indicates a “standard” distribution of dark matter though, some seem to have a lot less, or straightforwardly behave as expected without it. Often these are galaxies that have undergone collisions and the theory is that the dark matter was stripped away in the collision. MOND can’t account for these galaxies at all, hence MOND theories end up introducing ad hoc fields or versions of dark matter anyway.

Personally, the way I think about dark matter is that it’s a map of the variances we see between the expected behaviour and actual behaviour of galaxies. That’s it. That’s it’s a distribution of gravitating matter is one possibility. That would fit the observations, so it’s a reasonable working hypothesis, that’s all.

MOND simply doesn’t describe the distribution of the effects we see. It does in many common cases, but not all by a long way and the variance is different from case to case, so it still leaves a gap we’d need to fill with something else.

[throwawaymaths]

There's an n of 1 that is unexplainable by MOND, and that's the bullet cluster. Pretty much every other Galaxy is explainable using mond, the udgs for example are likely to be errors in distance to them, a lot of other errors due to being oriented face towards us, etc.

Iirc, LCDM has trouble explaining satellite galaxies. They're traveling through the halo, which should create drag on them. They shouldn't last long enough to exist at all.

[simonh]

For individual galaxies yes the bullet cluster is the main problem, but MOND can’t explain the motion of galaxy clusters at all.

[throwawaymaths]

> MOND can’t explain the motion of galaxy clusters at all

Yes that's fine, but it also doesn't necessarily try to [1]. LCDM can't explain linear Tully-fisher, can't explain external field effect, can't explain why satellite galaxies are long lived, can't explain why bullet cluster even exists (galaxies shouldn't be going that fast), can't explain early galaxies, etc.

[1] so maybe it shouldn't. There's a lot of systematic uncertainty in some of those really large scale movements. For example suppose our standard candle was off by half an order of magnitude at the largest scales. For LCDM, not a problem. Reparametrize the distribution and ratios of dark matter. Presto chango theory still works. That this should be possible should scare you.