Galaxy rotation curves don’t match Newtonian gravity. Stars at the edge of a galaxy orbit faster than the visible mass can explain. The standard fix is dark matter — an invisible mass halo that supplies the missing gravity. The alternative fix is MOND — a modification of gravity itself at low accelerations. Both approaches work. Both require multiple free parameters per galaxy. Both carry theoretical commitments that extend far beyond the data they explain.

A February 2026 model fits all 175 galaxies in the SPARC database — the standard benchmark for rotation curves — with a single free parameter. Not one parameter per galaxy. One parameter total. The model uses only luminous mass estimates and standard relativistic principles, with no dark matter halo and no modification of gravity’s force law.

The structural question this raises is about the relationship between fit quality and explanation. A model with one free parameter that fits 175 galaxies is parsimonious in a way that neither dark matter (which requires a halo profile per galaxy) nor MOND (which requires fine-tuning of the acceleration scale) can match. But parsimony is not truth. The single parameter might capture a deep physical principle, or it might be a well-constructed effective description that happens to span the data without touching the mechanism.

The through-claim: the number of free parameters is a measure of how much the model knows, but it is not a measure of what the model understands. A model with one parameter that fits everything has either discovered the correct generating principle or has found a mathematical surface that approximates the data without sharing its structure. The fit cannot distinguish these cases. The data are the same. The curves are the same. Only the physical story differs — and the story isn’t in the data; it’s in what the parameter means.

The galaxy rotation problem is not a shortage of models that fit. It is a surplus. The single dial adds one more fitting model to a field that has too many.