Lunar laser ranging to the Apollo corner reflectors measures the Moon receding from Earth at 3.82 plus or minus 0.07 centimeters per year, one of the most precise secular drifts in all of astronomy (Dickey et al. 1994; Williams et al. 2014). Run that rate backward with a constant tidal quality factor and the Moon sits at Earth's surface roughly 1.5 billion years ago, in flat contradiction with the 4.4-plus-billion-year age of the Earth-Moon system established by lunar samples and isotope chronology. The present recession rate is anomalously high, and the extrapolation paradox has shadowed tidal theory for a century.
The geophysical resolution requires the tidal dissipation rate to vary strongly over deep time: today's continental configuration happens to place the oceans near a resonance that amplifies tidal friction, while past configurations dissipated far less. Modern models thread the needle by tuning ocean-basin geometry against sparse deep-time proxies such as tidal rhythmites and Milankovitch sediment records (Laskar et al. 2004; Farhat et al. 2022; Auclair-Desrotour et al. 2022). The reconstruction works, but it is model-dependent, and the standard framework offers it no larger context: cosmology is assumed to stop at the solar-system door, so the lunar drift is a standalone geophysics problem with no relation to secular orbital evolution anywhere else.
The standing is a tension of frameworks rather than of data. The measurement is impeccable, the resonance models are increasingly capable, and yet the same pattern, bound systems slowly loosening, recurs at every scale from planetary orbits to galaxy clusters, and the standard model treats each instance as a separate puzzle with a separate bespoke mechanism.