Damage and Healing formula

[Masterkiller]

Here is an example of Force Charge

Force Charge (will look up rank and edit)
amountmodifier: -.39
coefficient: .91
Max tooltip 169
Max Weapon damage: 91
Bonus Damage: 39
X: 151

[(X * rank modifier) + Weapon + (Bonus * Coef)] * (1+amountmodifier)
[(151) + 91 + (39*.91)] * (1+-.39)
169

The same holds true with these:
with X = 151
Max tooltip | Max Weapon | Bonus

138 39 39
146.28 53 39
161.92 78 39
165.6 85 39

[Kaedis]

Problem is, with an empirically determined constant, there's zero reason to care about the actual formula. We know that there are 3 parts: weapon times coeff, bonus times coeff, and a static constant. If we have to empirically determine the static constant, might as well ensure accurate results by testing the coeff for weapon and bonus first. The only reason we should stick to a formula involving amountmodifier is if all of the data necessary for the formula can be pulled from a database. If we have to determine even one piece through empirical testing, simplifying the equation and double-checking each part makes more sense, particularly since we know the XML is out of date.

I'd love to know the actual equation the devs used, but we mustn't forget that the point of doing so is only so we have an accurate model of the way the abilities behave and scale at the levelcap.

Edit: Also, that doesn't really prove anything, MK. You need at least two datapoints on the same rank to demonstrate scaling, preferably 3 (where A and B are datapoints at the same bonus damage with differing weapon damage, B and C at the same weapon damage with differing bonus damage).

[metaspy]

(10-24-2011 04:20 PM)Freehugs Wrote:  Is anyone working on this stuff other than me?

I work on this stuff, but I am focused on the healing side.
I have to solely rely upon revealed info on levels of heal, willpower, and other factors that are revealed.

[Freehugs]

You're right MasterKiller, I read your formula wrong, because it does multiply weapon damage by a value.

(10-25-2011 12:03 PM)Kore Wrote:  If we're assuming an empirical test is needed to determine coefficient and base, the formula simplifies to WpnDmg*WpnCoeff + BonusDmg*BonusCoeff + C.

That's true, but only if Weapon Damage scales in a linear way, and I'm not convinced it does. I have a data set with the same melee damage bonus but 5 different weapons, and the results are not linear.

[Kaedis]

Quote:That's true, but only if Weapon Damage scales in a linear way, and I'm not convinced it does. I have a data set with the same melee damage bonus but 5 different weapons, and the results are not linear.

That makes so little sense I want to cry. Why would they make a non-linear weapon damage formula? It doesn't have to be 1:1, just linear. Anything else I would deem rather insane for the basis of an MMO damage formula.

[Masterkiller]

Freehugs can you post any of the non-linear datapoints?

[Freehugs]

You already have all my damage numbers, but for a simple example lets use Slash Rank 4 using the minimum damage numbers only.
Difference between:
Weapon 1 and 2: 8 damage
Weapon 2 and 3: 18 damage
Weapon 3 and 4: 4 damage
Weapon 5 and 6: 4 damage
Weapon damage numbers are displayed to 1 decimal place but always show whole numbers, so we'll assume that the weapons use the exact number listed.

Slash 4, difference between:
Weapon 1 and 2: 11 (10-12)
Weapon 2 and 3: 20 (19-21)
Weapon 3 and 4: 4 (3-5)
Weapon 4 and 5: 4 (3-5)
Tooltip damage amounts are displayed without any decimal places, and are very likely to have some fraction of a whole number so I've put up a damage range that allows for -/+0.5 on each tooltip value, giving a maximum variation of +/-1 when comparing tooltips. Thus the range in brackets.

So here's the problem, tooltip damage difference divided by weapon damage difference:
Weapon 1 and 2: Between 1.25 and 1.5
Weapon 2 and 3: Between 1.06 to 1.17
The first range does not overlap the second range, so there is no multiple of the weapon damages that matches all 3 tooltips.

7 of the the 18 abilities I've checked do not have linear progression with weapon damage. In all cases is the minimum damages that don't fit, but that's about the only pattern.

[Kaedis]

Thought: could the base damage be based on rank and weapon itemlevel rather than character level? That would cause non-linear progression like that.

[Ezmode]

That just sounds like redundancy to me, since you use the ilvl to get the weapon damage anyway.

[Kaedis]

True. Then again, I don't know of too many other ways of having a non-linear progression. If the damage formula is accurate, they've already got redundant with the coefficients and the amountmodifiers multiplied on top.

Could also throw out the assumption that the weapon tooltips are exact. Ie. they are listed to 1 decimal place of precision, but they actual round to an integer value before display. That actually kinda sounds just like an easy code bug, someone put a round(damage,0) instead of round(damage,1).

For the data points, if weapon 1-2 was actually 8.4, and weapon 2-3 was actually 17.5, it gives us almost exactly a scalar of 1.2. Then 2-3 and 3-4 could be 3.5 instead of 4, that would give 4.2 for the tooltip.