**Edit:** I accidentally checked “Set intersect to 0” in the original graph, which forced the equation to go through (0,0). After unchecking this, the updated equation becomes “y = 0.561x + 0.6283”, which is likely inaccurate as it would mean a calculated positive damage at 0 AP (of course, in-game the damage may round up to 1, but that’d be due to a forced minimum damage rather than the scaling). This indicates either the data being drastically effected by a margin of error (due to small sample size) or maehwa DP is still bugged (despite patch notes indicating otherwise) and the 166 DP isn’t actually reducing anything here. Will need to do some more DP testing against another class to confirm which it is. Either way though, it still seems that AP damage scaling is linear regardless of how DP scaling works.
So I did some AP damage scaling tests with my witch to see how AP affected damage dealt to players. For the test, I used my lvl 55 witch with up to 123 AP against a lvl 53 Maehwa with 166 DP, using Multiple Magic Missiles (which have 100% crit chance so we don’t have to worry about crits vs non-crits) with 123/102/88 AP and recording the damage dealt 70 times each (since that was how many trials it took to go through 300 HP pots). While the sample size isn’t all that big and 123 AP and 166 DP aren’t really large numbers, they were consistent, so I could use the data for determining the general scaling relationship (between AP and damage) even without knowing the actual damage formula.
Multiple Magic Missile hits up to 7 times, with most of my hits dealing either 3 or 4 hits. For each set of data, I estimated the average damage per hit, used that to determine the number of hits for each data point, calculated the actual per-hit average damage, then kept adjusting until the estimated average matched the calculated average. The end result was then compared against the AP value and graphed. The full spreadsheet can be viewed [here](https://onedrive.live.com/redir?resid=99464FFCFB5C54C2!1118498&authkey=!AM_LGtXWMcEMfEU&ithint=file%2cxlsx) for anyone who wants to cross-check the calculations/raw data.
The end conclusion is that AP damage scales linearly. At 88 AP, each hit of my Multiple Magic Missile on my volunteer test dummy does roughly 50 damage, while at 102 AP, its 57 damage and at 123 AP its 70 damage. The average damage per hit per AP values for the 3 sets were pretty close, and the R^2 value’s not too bad at 0.9953, so I’d say the data’s pretty accurate. Extrapolating the data, this tells me that at 100 AP, if I were to do 200 damage with my Multiple Magic Missile, I would do 300 damage at 150 AP.
Now of course, there are a lot of factors to consider. The level difference, the multiplier of the skill being used, the +/- RNG damage range, hidden stats, and much more. And of course the small sample size of only 70 trials. But the general scaling relationship still seems to be linear. Meaning that +10 AP will give you the same amount of damage boost at 100 AP as it will at 110 AP.
Will probably do some more AP damage testing later once I pick up some proper AP accessories, as well as some DP damage reduction testing. I really do wish this game gave more actual numbers though (or at least a damage log) to make things like these much less tedious…