Galaxy dynamics obey a law tighter than the model that explains them. The radial acceleration relation (RAR) links the observed centripetal acceleration in galaxies to the acceleration their baryons alone would produce, across 153 SPARC galaxies and 2700 individual points, with one characteristic scale, g-dagger = 1.2 x 10^-10 m/s^2, and an observed scatter of just 0.13 dex, consistent with measurement error alone (McGaugh, Lelli and Schombert 2016). Its cousin, the baryonic Tully-Fisher relation, ties total baryonic mass to the fourth power of flat rotation velocity with equally minimal intrinsic scatter. Galaxies of wildly different histories, gas fractions, and environments land on the same curve, as if dynamics knew only the baryons.
For ΛCDM this tightness is the problem. Rotation curves are supposed to reflect dark halos whose concentrations, spins, and assembly histories scatter stochastically, filtered through feedback-dependent galaxy formation; the predicted RAR scatter from semi-analytic models and hydrodynamical simulations is several times the observed value, and diversity should grow at low accelerations where halos dominate, exactly where the relation stays clean (Desmond 2017). Reproducing a zero-intrinsic-scatter baryon-locked law requires the halo and the baryons to conspire galaxy by galaxy, a fine-tuning of the very stochasticity the model is built on. MOND predicts the relation by construction but fails at cluster scales and in two-body tests.
The standing is foundational: the RAR is among the strongest empirical regularities in extragalactic astronomy, every new kinematic survey (including weak-lensing extensions of the relation to 100 kpc scales) confirms its tightness, and no consensus ΛCDM mechanism generates a law this clean from processes this noisy.