The Bullet Cluster cuts both ways: its 720 kpc gas-lensing offset is the particle picture's best exhibit (Clowe 2006), while its ~4700 km/s collision velocity sits at or beyond the extreme tail of what hierarchical formation produces (Markevitch 2004), improbable enough to be wielded as a ΛCDM problem in its own right. The photograph embarrasses its owner even as it convicts the alternatives.
The model needs the offset to prove collisionless particles and the velocity to stay within merger statistics, and the two demands pull apart: the cleaner the separation, the more violent the collision, and the violence is what the framework's gentle hierarchical assembly cannot afford.
SCT answers the two halves with registered machinery. The velocity half is the framework's home turf: extreme collision velocities are generic products of the cascade's kinematic hierarchy rather than tail events of gentle assembly (P22, P36), the same reading that absorbs El Gordo's improbability, so the Bullet's speed is an expectation, not an embarrassment. The offset half: lensing mass in SCT is amplified baryonic mass, Φ_local plus the coherent mesh term (P50, P52), and in a violent collision the components decohere differently, the shocked gas is kinematically stirred, suppressing its coherence amplification, while the galaxy ensembles pass through cold and comoving, their amplification intact, so the effective lensing mass leads with the galaxies even though nothing collisionless is present.
The honest accounting is registered: Paper 13, From Chaos to Coherent Gravity, computed a preliminary offset of 390 kpc against the observed 720, a factor 1.8 short, and the canonical ledger holds the full SCT merger simulation, Bullet initial conditions at production resolution, as the open task that decides whether the coherence-differential force law closes the gap. SCT claims the mechanism, not yet the number.
PARTIAL. The preliminary SCT estimate reaches 390 kpc of the observed 720 kpc offset. Resolution awaits the registered full merger simulation; until it runs, the Bullet remains the framework's most important unfinished calculation.
The registered kill is the simulation itself: if the full SCT merger run, with the coherence-dependent amplification implemented, produces an offset bounded well below 720 kpc under any admissible parameters, the M6 reading of the Bullet fails and the system reverts to particle evidence. The velocity half cuts at ΛCDM instead: merger statistics confirming the Bullet and El Gordo as compatible with hierarchical assembly would remove the speed embarrassment SCT explains.