Dragonfly 44 was announced as 99.99 percent dark matter: dwarf-mass stars, a 33 km/s dispersion, and 80 to 100 globular clusters implying a Milky-Way-class halo. The recount (Saifollahi et al. 2021) finds about 20 clusters, normal for a dwarf, leaving the kinematic excess contested. Either ΛCDM owns a dwarf living in a giant's halo, breaking its stellar-to-halo-mass relation, or its GC-count mass ladder is unreliable across the entire UDG census.
The model ties everything to one halo number: stellar mass, dispersion, and GC counts must all trace it consistently. UDGs spanning DF2-class emptiness to DF44-class excess force the halo to be absent and oversized within one morphological family, and no single ΛCDM formation channel covers that range.
SCT replaces the halo with coherence: apparent dynamical mass is baryonic mass amplified by A, governed by the coherence function C = exp(-σ_v²/v_cross²), which runs continuously from A near 1 in decohered systems to the virialized fixed point A* = 5.970 in fully coherent ones (P50, P52, P53). The UDG family then spans a predicted spectrum, not a paradox: DF2-class systems are tidally decohered (A of 1, no dark signature), while DF44-class systems are cold, coherent, and amplified, their elevated dispersions reading as large apparent halos without any actual giant halo to explain. One C-function covers the range that needs two ΛCDM channels.
The GC dispute matters less here than it does for the standard model, because SCT's dynamical reading rests on kinematics and coherence state rather than on GC-count halo calibrations: a DF44 with 20 clusters and a genuine 33 km/s dispersion is simply a coherent UDG sitting toward the amplified end, no Milky-Way halo required. The framework's registered population law does the discriminating work: across UDGs, A_obs must fall monotonically with σ_v/v_bulk, placing DF44 and DF2 on opposite ends of one measurable curve (P54).
This is the same coherence physics behind the DF2/DF4 deficiency and cluster amplification. There is no need to invoke a separate failed-giant formation channel.
The registered kill: across twenty or more UDGs with measured kinematics, no monotonic inverse correlation between A_obs and σ_v/v_bulk at more than 2σ falsifies the coherence spectrum reading. DF44 itself is a lever: if improved kinematics confirm a dwarf-normal dispersion alongside the dwarf-normal GC count, the system exits the tension entirely; if the 33 km/s excess holds while the system shows high internal coherence, it must sit on the predicted curve or the mechanism fails.