Growth Rate mechanism revisited

[Yay]

Hello,

I wanted to check and adjust Growth Rates for many classes and stumbled across, what I thought was, a small inaccuracy. I redid my tests more thoroughly and came to the conclusion that the growth rate calculation misses something. To be sure that it is handled like this I checked on Storm then and found it working exactly like it is described in Wyrd's Spreadsheet (that was likely correct at that time). Problem is, that this seems no longer correct.

Good News is, that there is still the Growth Rate mechanism, but on top of that there is an offset that doesn't seem to have another dependency. I think it's another value that is related to this style. This offset is quite low compared to the impact of the Growth Rate at level 50, but it's important to lower level styles and also to the Growth Rate calculation in general, because the values seem mostly only correct at level 50 (sth like +11 and such).

The damage multiplier/ratio for styleCap/unstyledCap is the same for hits below cap. This makes testing way more consistent. Also DPS of the weapon doesn't matter for testing. The styleDamage per second is the same regardless of the weapon DPS (if you transform the formula from Wyrd that is to be expected). This also applies to the Offset, so the Offset behaves like a DamAdd (but only if the style is performed successfully ofc).

Before I propose a new formula; here is the setup and my test values:



Values with blue background are the trained WeaponSpec (without item bonuses). These are styles from the Albion Crush line (Paladin).

GrowthRate is interpolated with the first and last values divided by the difference of WeaponSpec like this: (DamageAt50-DamageAt10)/(50-10).
Absolute Offset is extrapolated with the GrowthRate, so the Weaponspec is theoretically 0.
NormalizedOffset is divided by effective weapon speed.










It's obvious that the linear growth is not beginning at the unstyled Cap, but at an offset that is only influenced by the swing speed (L8 Bludgeon might be an exception).
I also tested some extremes with level 11 character and no spec (L1 Style) and such things, there is nothing weird happening, only the usual fluctuations.

My proposed formula would be like this:

absoluteStyleDamage = baseDam * ToA-StyleDam + (baseDam/unstyledCap) *(normalizedOffset + GrowthRate * WeaponSpec) * effWeaponSPD)

Where GrowthRate is essentially damage per second per WeaponSpec and normalizedOffset is damage per second.
EffWeaponSPD is calculated as usual: WeaponSPD*(1-(ToA-MeleeSPD + Haste + Celerity)/100)*(1-(Qui-60)/500).

(Note: ToA-StyleDamage is only there for completeness. It seems to be unaffected by Haste. Tested quickly with Detaunt. )

[Leodagan]

Pretty much interesting

And what could be the normalizedOffset ? something like ratio * GrowthRate ?

Maybe we lack some "WeaponSpec + 1" * GrowthRate -- or -- "WeaponSpec + ratio" * GrowthRate ?

[Yay]

As far as I see it, there is no correlation to GrowthRate and normalizedOffset. It's only influenced by absorb (or baseDam/unstyledCap) and SwingSpeed (effWeaponSpeed). It acts like an offset to the linear GrowthRate function.

Look at L1 (Daze) and it should be obvious that there is really no GrowthRate (=0) at all, but it does style damage. L8 (Bludgeon) on the other side is at least very close to a linear growth without offset (I guess it exists, but is very small). Every other style in this test has an offset that's big enough to mess up calculation.

I think the normalizedOffset (or whatever you wanna call it) is unique to each style. My only relation I could think of is the style level, but the function would look weird (abs or some kind of parabola function maybe). I can't completely rule it out as of yet, but I really don't think there is a correlation.