Could baryonic feedback plus anisotropy together over-fit the core-cusp problem without revealing new physics?

This is a sharply posed question that cuts to the heart of a real methodological tension in dark matter astrophysics.

The short answer: the archive suggests the answer is "not cleanly" — baryonic feedback and anisotropy can suppress the cusps, but they appear to leave detectable signatures that distinguish them from true new physics.

What baryonic feedback can do

Baryonic feedback (supernovae, AGN jets) does reshape dark matter profiles. [AG-2025.01-276] shows in FIRE-2 simulations that stellar feedback in cold dark matter (CDM) can create cores in dwarf galaxies, converting cusps into flat central densities. This is a real effect, not a numerological artifact. However, the key finding is that feedback-induced cores have a specific signature: they transition relatively slowly from the central plateau to the falling density at larger radii, compared to other mechanisms [AG-2025.01-276].

The reason this matters: if you only had rotation curve data, you might fit many profiles. But once you add stellar kinematics and density distributions—*which constrain the shape independently*—you gain leverage to distinguish causes.

The over-fitting problem

You're right to worry about over-fitting. [AG-2025.02-271] demonstrates that a purely statistical model (DARKexp) can fit 96 galaxies in the SPARC catalog reasonably well with a single shape parameter, even without feedback. This suggests that shape flexibility alone is not enough to falsify CDM plus feedback.

But the paper also shows something interesting: DARKexp fits both rotation curves *and* the projected stellar density of ultrafaint dwarfs [AG-2025.02-271]. That's two independent datasets, not one.

The smoking gun: ultra-faint dwarfs

Here's where the archive gets pointed. [AG-2024.07-1420] argues that six ultra-faint dwarf galaxies (UFDs, with stellar mass 10³–10⁴ M☉) show stellar *core* distributions that are inconsistent with CDM potentials *even under spherical symmetry and isotropic velocities*—and crucially, their stellar mass is too low for feedback to have reshaped the dark matter potential. Using Eddington inversion, the observed innermost slope of stars rules out CDM at >97% confidence.

This is the key point: anisotropy and feedback become excuses only if you have enough baryons to do the work. UFDs don't.

Alternative: fermionic degeneracy

[AG-2024.07-1083] proposes a different resolution within CDM: degenerate fermionic dark matter (not exotic particles, but a statistical effect in fermionic halos) can naturally produce cores without feedback. The mechanism is that degeneracies in small subhalos suppress density in the host halo over large scales. This is still CDM in spirit, but it produces the diversity problem without fine-tuning feedback.

The bottom line

Baryonic feedback plus anisotropy can reduce tension with observations, but the archive indicates they cannot fully explain away the problem without creating new tensions:

• Feedback works for larger galaxies but fails (or requires implausible tuning) in ultra-faint systems [AG-2024.07-1420].
• The density gradients of feedback-induced cores are observationally distinguishable from self-interacting or degenerate dark matter cores [AG-2025.01-276].
• Once you fit rotation curves, you must also match the radial stellar distributions and halo baryon fractions across multiple independent probes—and the archive suggests this joint fit is harder than it looks [AG-2025.01-296].

In short: feedback is real and important, but over-fitting one observable (rotation curves) while ignoring others (stellar kinematics, baryon distribution) would be methodologically unsound. The archive points toward genuine tensions that require either new dark matter physics or a more sophisticated model of feedback—not just parameter tuning.