The load-bearing part of the new Yonsei paper is arithmetic rather than astronomy. The age-bias correction is a product of two quantities: how much supernova progenitor age changes with redshift, and the slope of the age-luminosity relation [9]. Those factors are not independent. Because the age difference sits in the denominator when the slope is fitted, shrinking the assumed age evolution necessarily steepens the slope, and the two moves largely cancel [10]. Halve one factor and you roughly double the other, leaving the product that actually gets applied to the distances close to where it started [13]. Treated consistently, Yonsei says, the correction and the cosmology that follows from it come out much as before [11]. On that reading the rebuttal's second argument was never about the correction; it was about one of its two factors.
The first argument is more testable. Measuring the age-luminosity relation requires comparing supernovae inside a narrow redshift interval, or population evolution across the range gets entangled with the correlation being measured [15]. The rebuttal pooled an unusually broad interval instead [6], and Yonsei reports that the recovered slope gets systematically shallower as higher-redshift objects are added, in mock data built to mimic the observing conditions as well as in the real sample [7]. That is a claim with a shape another group can reproduce or break without new observations. "When the comparison is restricted to an appropriate redshift interval, the strong age dependence reappears," says lead author Chul Chung [8].
The dust criticism is separate and blunter: Yonsei says the extinction model behind the rebuttal needs a host-galaxy mass dependence with the opposite sign to what large galaxy samples show, so that piece of the model has no observational support [12].
Worth noting where all of this comes from. The account of the rebuttal's defects is the Yonsei team's own, relayed by phys.org [1]; the Southampton-led group's response to it is not in the material. That asymmetry weighs differently on the two arguments. The denominator point is algebra and can be checked on paper. The slope-flattening trend depends on sample cuts a reader cannot inspect from a summary.
What would actually close the question is an estimate of progenitor ages that does not inherit its redshift evolution from host-galaxy ages, since that inheritance is exactly what the June paper disputed [c5c]. Absent that, the two rounds of argument, roughly seven months apart [14], have moved the dispute off whether an age bias exists and onto how its slope should be measured. The word "crisis" in both write-ups [16] is doing work that the fitted numbers have not yet done.
