The best way to test MiHsC (quantised inertia) is to look at systems where accelerations are very low, and so the anomalies it predicts become more obvious. Milky Way satellite dwarf galaxies are brilliant tests, being far from the Milky Way and having only a tenuous hold on their stars.
It is well known that full-sized galaxies spin too fast to hold themselves in gravitationally. So astrophysicists add extra invisible (dark) matter to them, putting it where they want, in a deeply unscientific manner. These satellite dwarf galaxies are useful because it is very hard for them to do that, because dark matter usually has to stay spread out on huge galactic scales to explain why it remains only around the edge of galaxies, so you can't then suddenly pack it into a tiny dwarf galaxy, without causing a contradiction. Also, the amounts of dark matter needed to hold these dwarf galaxies together is 100 or more times the visible mass in them which is getting to be ridiculous, especially given the fact that you can't concentrate the stuff like this anyway without becoming intellectually schizophrenic.
I have recently compared the predictions of the dwarf galaxies rotation speeds from MiHsC (which reduces the inertial mass for low accelerations in a new way), the speed predicted by Newtonian or general relativisitic models (without dark matter) and MoND with the speeds observed. This plot is the result:
The observed spins (velocity dispersions of the stars) of these 11 dwarf galaxies (the ones for which I can find both mass and velocity data, I have not cherry picked) are shown by the hollow squares. The dwarfs' names are also shown. The predictions of Newton/GR (without dark matter) are shown by the crosses at the bottom. The predictions of the empirical theory MoND are shown by the black triangles (with its adjustable parameter set at 1.8x10^-10 m/s^2) and the predictions of MiHsC are shown by the black diamonds. The root mean squared error for Newton, MoND and MiHsC are 6.3, 3.3, 2.9 km/s respectively, so MiHsC is the closest to the observations, despite having no adjustability, and in these cases applying dark matter is doubly ridiculous, as mentioned above.
The dark matter detection industry is unlikely to be happy about this, but the point is you can make things work out with MiHsC on a piece of paper without spending millions on huge detectors. This deserves some notice. I'm just about to submit a paper so feedback or suggestions for more data would be very welcome.