Jerebear wrote:(base damage - S1 - S2 - ....) * (B1 + B2 -D1 - D2 + ....)
The problem is that in this case, the shield has to multiply to determine how much damage it should subtract. It's doing that before percent-based modifiers, thus it's not shielding 50% of the actual damage, it's only shielding 50% of the pre-buffed damage.
Shields (as I am referring to them) don't reduce damage by a percentage but by a fixed value. But for the sake of semantics, in the equation I mentioned
S1 and S2 are fixed damage reductions (stoneskin for example)
B1 and B2 are percentage based increases (shattered defenses for example)
D1 and D2 are percentage based decreases (crouch for example)
Looking back, I picked those letterings poorly.
So I think we are talking about two different things.
There are also some other factors that I didn't include:
Fixed Percentage damage (explode, etc.)
Fixed damage buffs (wild magic, etc.)
I haven't exclusively tested wild magic, but it should be included in with the S1 and S2's area (though as a + rather than a -). Fixed percentages like explode are just tacked on outside of any modifiers.
To alleviate (hopefully) any confusion, I'll relabel the equation:
Base = damage the attack normally does
FV = Fixed Percentage Attacks (Explode, etc)
FD = Fixed value debuff (wild magic, etc.)
FB = Fixed value buff (stoneskin, etc.)
PD = Percentage value debuff (shattered defenses, etc.). Values in decimal (50% is 0.5)
PB = Percentage value buff (crouch, etc.). Values in decimal (50% is 0.5)
FV + [(Base + FD - FB) * (1 + PD - PB)]
Though I haven't tested FD to verify that. So if you talk about a 50% "shield" it will be in the PB catagory. Everything that I have tested seems consistent with that model. If you have multiple PB's they can be in any order (same goes for any of the variables...order doesn't matter because they are grouped associatively)