The way he described it in the bug report, he took the ratio of damage at 5 APS to the damage at 2.13 APS? But if it’s a flat scalar, then it appears in both damage components. So it cancels out.
E.g. if damage5 = base damage * 5 * (unknown scalar)
and damage2.13 = base damage * 2.13 * (unknown scalar)
damage5 / (damage2.13 / 2.13) = (base damage * 5 * unknown scalar) / ((base damage *2.13 * unknown scalar)/2.13) = 5, i.e. the ratio will be 5 regardless of whether there is an APS scalar in the damage calculation, and regardless of what it’s value is.
Based on what he said in the bug report, I think a flat scalar can explain what he observed.
Or when he tested it live was it more involved than that?