> What can't be fit by declaring the amount of dark matter that must be present fits the data?
Tons of things - just like there are tons of things that can't be fit by declaring the amount of electromagnetically-interacting matter that must be present fits the data.
You can fit anything you like by positing new and more complicated laws of physics, but that's not what's going on here. Dark matter is ordinary mass gravitating in an ordinary way: the observed gravitational lensing needs to match up with the rotation curves needs to match up with the velocity distributions of galaxies in clusters; you don't strictly need large scale homogeneity and isotropy but you really really want it, etc. Lambda-CDM doesn't handle everything perfectly (which in itself demonstrates that it's not mindless overfitting) but neither does anything else.
There are modified gravity theories that are compatible/extensions to GR, e.g the f(R) gravity theories.
Nobody probably believes MOND as such is some fundamental theory, rather as a "theory" it's sort of a stepping stone. Also MOND is used often interchangeably (and confusingly) with modified gravity theories in general.
> Dark matter is ordinary mass gravitating in an ordinary way: the observed gravitational lensing needs to match up with the rotation curves needs to match up with the velocity distributions of galaxies in clusters
Those are all the same thing, the shape of spacetime. The only thing DM adds is a backstory that this shaping comes from hypothetical undiscovered particles with properties that match observations.
Tons of things - just like there are tons of things that can't be fit by declaring the amount of electromagnetically-interacting matter that must be present fits the data.
You can fit anything you like by positing new and more complicated laws of physics, but that's not what's going on here. Dark matter is ordinary mass gravitating in an ordinary way: the observed gravitational lensing needs to match up with the rotation curves needs to match up with the velocity distributions of galaxies in clusters; you don't strictly need large scale homogeneity and isotropy but you really really want it, etc. Lambda-CDM doesn't handle everything perfectly (which in itself demonstrates that it's not mindless overfitting) but neither does anything else.