I can't stay up any longer.
Here's my raw data, such as it is.
I recorded 99 non-crit melee hits with Serenity as WHM/BLM, and used this as a baseline for RDM/BLM (DL I) and RDM/DRK (DL II). Unless I'm spreadsheeting wrong, which is possible at this hour, for DL I and DL II
WHM Baseline : 841.3 (100%)
RDM/BLM DL I : 863.5 (102.6%)
RDM/DRK DL II: 886.7 (105.4%)
I recorded 99 non-crit melee hits with Fettering Blade as COR/BLM, and used this as a baseline for DRK/BLM (DL V).
COR Baseline : 519.2 (100%)
DRK/BLM DL V : 598.7 (115.3%)
I recorded 99 non-cirt melee hits with Atoyac as COR/BLM, and used this as a baseline for RNG/BLM (DL III). I have only collected 51 samples for RNG/BLM for DL III as I can't keep going, but here's what 51 samples says:
COR Baseline : 381.8 (100%)
RNG/BLM DL III: 417.8 (109.4%)
So, tentatively, percentages and possible /1024 if those are what they actually are:
Damage Limit I : +2.6% (27/1024)
Damage Limit II : +5.4% (55/1024)
Damage Limit III: +9.4% (96/1024)
Damage Limit IV : ???
Damage Limit V : +15.3% (157/1024
These aren't what I was expecting, and I only have 99 samples for DL I/II/V, and 51 for DL III, so they could be off. I took great care never to stab a bat when using Atoyac. And now I want to sleep hopefully someone who has the patience that only a machine can have will somehow find it in their heart to collect a much larger sample size for these.