The Hubble constant — the rate at which the universe is expanding — has been measured two ways. From the early universe: observe the cosmic microwave background, apply the standard cosmological model, and predict what the expansion rate should be today. From the local universe: measure distances to galaxies directly, observe their recession velocities, and calculate the rate. Published in Astronomy & Astrophysics, the H0DN Collaboration — 117 authors led by Stefano Casertano — combined twelve independent distance indicators into a single network and measured H0 = 73.50 ± 0.81 km/s/Mpc. One percent precision, for the first time.
The early-universe prediction is about 67.4 km/s/Mpc. The local measurement is 73.5. The difference — roughly 9% — is now 7.1 standard deviations from the Planck+SPT+ACT cosmological model. Five standard deviations is conventionally treated as discovery-level significance. Seven is beyond dispute as a statistical anomaly.
The structural insight is about what precision does to disagreement. The expectation was that better measurements would resolve the tension — that one method would converge toward the other as error bars shrank. Instead, better measurements sharpened the tension. The 1% measurement didn't explain the gap; it made the gap harder to explain away. Every improvement in precision has moved in the same direction: confirming that the two numbers genuinely disagree.
The robustness property deepens this. The network combines parallaxes, eclipsing binaries, masers, Cepheids, the tip of the red giant branch, Miras, JAGB stars, Type Ia supernovae, surface brightness fluctuations, Type II supernovae, the Fundamental Plane, and Tully-Fisher relations. No single measurement or method is critical to the result — any component can be removed entirely, and the value remains essentially unchanged. The disagreement is not an artifact of one measurement technique. It is a property of the universe as measured by every available method simultaneously.
This is the rare case where the answer to a scientific question gets worse with better data. Not worse in the sense of less accurate — worse in the sense of more troubling. The two measurements are both correct. They disagree. Something in the standard cosmological model is incomplete, and improved precision has narrowed the space where the missing physics could hide without changing any of the numbers.