Reusing an old parameter set is an underrated move in simulation work, and it is what makes this readable as evidence rather than as one more possible history. The comparison is narrow but clean: hold the configuration at the values the original giant-impact modeling used, down to equal internal temperature structures in both bodies, switch on temperature-dependent geologic strength, and see whether the outcome class moves [6][9]. According to the study team, it moves [11]. The impact code itself came from methods built at the University of Bern and the University of Arizona [14].
The mechanism the authors put forward is thermal, not chemical. Hotter rock is weaker rock, and the team reports that Moon formation is sensitive to the temperatures of the two colliding bodies [7]. That gives one dial with two outcome families hanging off it: a whole Moon within hours, or a protolunar disk that has to assemble a Moon afterwards [8]. The paper's title names the branch it cares about, calling the intact case a collisional capture that depends on strength [2].
Twenty-five years separate Canup and Asphaug's 2001 paper from this one [17], and across that span strength was left out of giant-impact models on the reasoning that it could not matter at such high energies [4][5]. That was an argument about magnitude, and it was a defensible one; Denton's group had previously found the same term mattered for asteroid-scale collisions and for Pluto-Charon [16], where nobody disputes that geology survives the event.
What the result leaves open is which thermal state Earth and Theia actually had. The dating argument runs through an inference: young protoplanets start hot and cool with age [10], so if hotter means weaker [7], the intact branch points toward a later, cooler collision than the disk branch does. Canup, who was not involved in the work, suggests the Moon's physical properties today, perhaps including its volatile content, could be read back to the thermal state at impact and used to constrain the timing [12]. That remains a research programme waiting to happen, and no one has run the measurement yet. An intact Moon likewise leaves the oldest complaint about the giant impact exactly where it stood: the near-identical composition of Earth and Moon, which the authors leave unresolved, alongside the suggestion that Theia and proto-Earth grew from a common region of the disk [13].
Two numbers I would want before treating the five-hour case as anything more than a demonstration: how many runs the parameter sweep contains, and where in temperature the boundary between capture and disruption sits. The published summary reports the two outcome classes without either [18]. My read is that temperature-dependent strength has become an input a giant-impact paper has to declare rather than assume away, because the outcome class was riding on an assumption that was not being reported [5][11].
