At minimum this is a ~200 billion parameter model, and more if you’re looking at smaller structures.
With dark matter it’s two steps removed where we’re inferring the behavior of baryonic matter and then inferring the amount of baryonic matter we aren’t observing and then calculating the existence of dark matter to get that behavior after accounting for undetected baryonic matter.
So there’s no simple way to calculate it from say looking at the Milky Way alone and extrapolating from the baryonic mass of the rest of the universe. Trying to approximate things from a representative sample is its own problem.
1. Dark matter does not meaningfully introduce billions more parameters into cosmological models than they already have, and
2. Individual galaxies' dark matter fractions are not essential to (not proving, but) strongly suggesting dark matter exists.
If the observed dark matter fractions of all known galaxies were 0% but the CMB was unchanged we wouldn’t assume dark matter exists. Thus your #2 is false. There’s infinite models consistent with any observation so finding something after a model was created for other reasons is useful as validation, but the chain of logic is still dependent on the prior observations not the model.
In a meaningfully different cosmos different observations would have happened and different models would exist. Trying to pick out specific experiments as sufficient on their own glosses over that particular limitation.
No, astrophysicists would eventually figure out something was up when they couldn't replicate the actual spectrum with dark-matter-free simulations. Why would you assume otherwise? Unless you want to dig into the assumptions of the scenario, in which case you're probably proposing a self-inconsistent universe so of course you can draw whatever conclusions you want from it.
> There’s infinite models consistent with any observation...
You can't actually believe this and still believe in science. If observations don't constrain models, then there is no point in observing. And in the long run, there's asymptotically no difference between "prior observations" and later observations. They're just observations that all go into the same model-constraining mill. Scientists are not fools, and are capable of realizing when an initial observation put them on a wrong trail.
You're still barely touching the real point. This all just sounds like rationalizations to avoid the fact that dark matter, for now at least, and for all that it genuinely sucks, is the Occam's razor explanation for the full suite of observations. Why is this so hard to accept?
The point is that even with current observational data there's no reasonable distribution of dark matter that correctly explains all evidence that we have.
Your intuition that "if I have an infinite number of degrees of freedom anything at all can be fit" is leading you astray here.
This obviously can match almost anything and it has extremely low predictive power (many future observations may differ from predictions, which can be accounted by some dark matter whose distribution was previously unknown), so it is a much worse explanation than a modified theory of gravity that would have only a finite number of additional parameters.