Can absolutely confirm Armor Shatterer as having a 50% Dex mod. Removing that leaves the base+fStr damage value on str-capped AS as solely dependent on skill. EG: At 427 skill and capped fStr, base damage will be 154 before Dex WSC is added in, regardless of Dex.
Don't Sylow's Hpmede tests contradict this?
Code
Control : Thunder Maneuver
Level 73 : 76 VIT : 3255 : 3300
Level 74/75 : 77 VIT : 3240 : 3285
Level 76/77 : 80 VIT : 3225 : 3285
If it's just a 50% DEX mod, why does adding DEX while keeping all other variables constant eliminate a VIT-based damage decrease?
Hmm. Honestly, I'm not sure how to explain that, regardless of the dex mod. It flat out shouldn't be possible, assuming other factors remained constant.
For each increase in mob vit, control damage went down by 15 (and thus base damage dropped by 1). A thunder maneuver should add a fixed amount to dex, and thus a fixed amount to the final damage. For the first two vit values, damage increases by 3 base damage each, which is exactly as predicted (166 dex => 70 WSC, 173/174 dex => 73 WSC). [Note: It's not indicated whether those Maneuvers are basic +7 stat or enhanced +8 stat, but it works out either way.] For the last one, damage increases by 4 base damage.
We can't factor in some sort of limiter on the dex based on fStr because doing so would imply that the final value should be -lower-, rather than higher. We can't say that dex increases the weapon rank, thus allowing higher max fStr, because str remained constant between the two data set, and we weren't capped on fStr to begin with (as evidenced by the Fire Maneuver samples).
Unfortunately there aren't any AF3+2 + Thunder Maneuver data points to compare with on the highest vit targets. For the lower vit targets, the difference is 4 base damage, which again is expected (179 dex => 75 WSC, 186/187 dex => 79 WSC).
Analyzing the lowest level sample:
76 vit target, 3255 damage == 217 base
Assuming 427 skill (not specified on spreadsheet), base damage from skill is 92+x
Dex WSC is 166 * .5 * .85 = 70
Str is 162; dStr is 86; fStr2 is 45
Total base damage is 92 + x + 70 + 45 = 207 + x
207 + x = 217
x = 10
That matches the fixed base value implied by my test data as well.
Analyzing the vit 80 target:
80 vit target, 3225 damage == 215 base
Assuming 427 skill (not specified on spreadsheet), base damage from skill is 92+x
Dex WSC is 166 * .5 * .85 = 70
Str is 162; dStr is 82; fStr2 is 43
Total base damage is 92 + x + 70 + 43 = 205 + x
205 + x = 215
x = 10
Everything is a proper match, again.
Now the Thunder Maneuver vs vit 80:
Actually, the math isn't going to show anything different from what we already know, but I did notice something interesting. Vit 80 target with TM has exactly the same damage as Vit 80 with FM, which is the only instance of that being the case in all the samples. It's also exactly the same damage as that seen by the vit 77 targets with TM above it in the column.
At this point I'd expect a high likelihood of human error, of not completely entering the new value, and Excel copying the value that matched it from earlier in the column after typing '32', since the only values above it were 3300 and 3285. If he allowed the auto-complete to take over, that would explain the problem.
I'll see if I can test this.