MOND does work a lot better than Lambda Cold Dark Matter (L-CDM)... but only for explaining the rotation of galaxies. It can explain this rotation almost perfectly given only the mass distribution: no tuning is required. Dark matter, on the other hand, has issues with dwarf galaxies and has no predictive power: you just fit the dark matter to the results you observe (Which is why you end up with some dwarf galaxies that are almost entirely composed of dark matter, and ones that almost have none. With MOND it just works).

Of course MOND can't really explain the third peak of the cosmic microwave background radiation, so it isn't perfect either. It is also phenomenological, with no underlying physical theory at the moment. Still, it's surprising that it does work at all.

I should also mention that the bullet cluster, which is touted as proving dark matter, causes issues for dark matter as well as MOND. The velocities involved in the collision are higher than can be explained by the current dark matter models. MOND kind of sucks at dealing with clusters, as well.

TL;DR MOND is a lot better than LCDM at explaining galaxy dynamics, LCDM is a lot better than MOND at explaining the cosmic background radiation. Both aren't that great at dealing with clusters (but you can also make dark matter work with enough fiddling).

https://tritonstation.wordpress.com/2018/09/04/dwarf-satelli...