Yeah that's exactly (the latter part of) what I just said. PXP0 works in a complex way though, and also has a tiny effect in the opposite way (the first part of what I had said). Lately I've been trying to make progress understanding PXP0 better, so let's try to calculate the amount of tiny effect here.
QUOTE(The Wiki)
Derived_Base = (ceil( round(Base_Roll * Slot_Mod)) +
ceil( round(Suffix_Roll * Slot_Mod)) + Quality_Bonus) * Base_Multiplier
Quality_Bonus = (PXP0-100)/25
If an equipment is forged, the additional multiplier applies to Base_Roll and Suffix Roll only (i.e. Quality_Bonus part is unaffected)
I think the (overall) shittiest legendary rapier I've ever seen had 354 PXP0 and the (overall) best around 370 PXP0, so let's say the difference is 16 PXP0. This corresponds to a difference in Quality_Bonus of 10.8 - 10.16 = 0.64 times 0.0854 Base_Multiplier gives about 0.05 attack base difference from the PXP0 component.
Actually we should probably use the theoretical shittiest and best, so peerless 377 PXP0 and maybe 347 PXP0. So a difference in Quality_Bonus of 11.08 - 9.88 = 1.2 times 0.0854 gives about 0.1 attack base difference.
If these formulas are correct, which they probably are not, I guess this Hallowed Rapier could have had ~51.25 adb with an absolute perfect roll.