VaultNetwork.netVault Network Boards
Author Topic: Melee Speed, Melee and Style Damage (or why pure Grothrates are wrong!) [Locked]
Vanesyra  1 star
Posts: 108
Registered: 2009-12-8 17:31:44
first of all some infos:
* calculations i will provide here should be +/-1 of the real ingame values. daoc uses a lot of round down (double to integer) what makes it hard to get a exact calculation done.
* calculation is not working well (more then +/-1) for some "bugged" items. (items that show 3.2 speed if you "right click" them and if you "shift + i" them you get 3.3 speed, or if you are rr5+ and the game still shows you a "- 16.5 Base DPS (16.5 Clamped DPS)" on the "shift + i" of the item. "/bugreport" this items, they are really bugged not just because my calcs do not match perfect here ^^)
* if i use (3.3 + 4.4) = 7.7 the brackets are normal brackets
* if i use [3.3 + 4.4] = 7 the brackets indicate a round down
* if i use MIN(1, 4) the lowest number is selected from this list: 3 * MIN(4, 2) = 3 * 2 = 6
* when talking about "SPEC" i refer to: "spec points + item bonus + rr bonus" for example 50 + 11 + 12 = 73 => this is the SPEC value
* when using values like "HASTE", "CELERITY", "TOASPEED", "TOADAMAGE", "TOASTYLE" i am using them as 0. values. for example 20% haste is used as 0.2


melee speed:
CURRENTSPEED = [[[BASEWEAPONSPEED * 100 * (1 – (MIN(250, QUICKNESS) – 50) / 500)] * (1 – HASTE – CELERITY)] * (1 – TOASPEED)]

the game calculates up to 0.01 seconds. and yes its "QUICKNESS - 50" and "TOASPEED" is split away from "HASTE and CELERITY" this way. see http://camelotherald.wikia.com/wiki/Quickness#Melee


twohandbonus
TWOHANDBONUS = 1.10 + 0.005 * SPEC

if it's a two handed weapon and if its not it's "1.0"


unstyled cap damage
DAMAGECAP = [[[[[WEAPONDPS * BASEWEAPONSPEED * 10 ] * (0.94 + BASEWEAPONSPEED * 0.03) + MYTHICALDPS] * TWOHANDBONUS] * TOADAMAGE] / 10 * 3]

a side note here, with the way how mythical dps works. it is better to use a mythical dps myth on a one handed weapon instead of a +8 str myth. and it's better to use a +8 str myth on a slow twohanded weapon.


style damage
STYLE DAMAGE = (BASE + SPEC * GROTH) / 10 * CURRENTSPEED * UNSTYLEDDAMAGE / DAMAGECAP + UNSTYLEDDAMAGE * TOASTYLE

BASE and GROTH are new values. a single grothrate will not work here. both are integer values. you can find some example values for them at http://camelotherald.wikia.com/wiki/Two_Handed and http://camelotherald.wikia.com/wiki/Crush in the "Damage" cell. there are 3 values: the first is the calculated (BASE + SPEC * GROTH) for a rr 13 +11item 50 in weapon char (for you to compare the styles), the 2. value is the BASE and the 3. value is the GROTH of the style.

very interesting is to see that the damage from the style is depending on the current weapon speed. so for example if you swing a 6.0 weapon at 6.0 speed and you deal +100 style damage then you will deal +50 style damage if you swing the same weapon with 3.0 speed. but the toa style damage bonus is different! this is a flat bonus and does not depend on your current weapon speed. so when ever you do a template, never ever drop toa style damage. you are going to loose a huge amount of damage if you do so. for example the +10% toa style damage on my paladina increase the over all style damage by 50%, this is huge!

this shows us that a pure grothrate does not do it. we need a base value and a groth factor to do it. it also shows us that toa style damage is something you should never drop.

up to now i only found "GROTH" values: 0, 3, 6 and 9 where in most cases "Damage: very low" and "Damage: none" => 0, "Damage: low" => 3, "Damage: medium" => 6 and "Damage: high" => 9. the "BASE" values are pure integer values, positive and negative once and i did not find a pattern here yet.


not every thing is new stuff here, but some things should be new sorry for this english, hope someone can make a use of this calculations anyway or even better find errors ;P
Thak0r  1 star
Posts: 170
Registered: 2009-7-21 02:56:50
pretty interesting.

 

-----signature-----
Thakirion, Eld, 11L4 | Troikhan, BM, 11L5
Wulfher, Hero, 11L1, Lone Enforcer
Rahjan, VW, 9L2 | Thakisa, Ment, 9Lx
Thumb, Ranger, 8Lx | Ignitez, NS, 7Lx | Thallak, Warden, 7Lx
http://gimpchimp.etilader.com/s.php?u=Thaki
DoorknobMLF  3 stars
Posts: 627
Registered: 2008-3-2 09:16:57
If I played melee classes a lot, and had a lot of time, I would do something like this too, since I have the math and statistics knowledge to do these tests. But there could be a lot of factors acting in many unknown ways, so it could take a very long time to get it right.

These formula's might approximate damage quite well, but i'm no convinced that they are whats actually used in the ingame calculations.

Btw I'm curious, where did you get this info?

 

-----signature-----
UnscrupulousDulu  1 star
Posts: 152
Registered:
DoorknobMLF posted:

If I played melee classes a lot, and had a lot of time, I would do something like this too, since I have the math and statistics knowledge to do these tests. But there could be a lot of factors acting in many unknown ways, so it could take a very long time to get it right.

These formula's might approximate damage quite well, but i'm no convinced that they are whats actually used in the ingame calculations.

Btw I'm curious, where did you get this info?



"If I played melee classes a lot, and had a lot of time, I would do something like this too,"


First I was all like:


"but i'm no[sic] convinced that they are whats actually used in the ingame calculations."


Then I was all like:


@ Vanesyra - Excellent work.
NeehiK
Title: Perma-newb
Posts: 49
Registered: 2002-10-20 16:18:46
Sorry, it's killing me... it's groWth. Very interesting calcs.

 

-----signature-----
Thalatos
Sorcerer on Port Nowhere
Windwalkr  1 star
Title: Camelot Vault Staff
Senior Mentor

Posts: 180
Registered: 2002-7-26 11:47:42
Looks like quite possibly the most important calculations discoveries --related to melee mechanics-- since Wyrd spec. (Provided it all turns out right, but I sure as heck won't be the one to tell you it's not. )

Excellent work, I will let some more experts weigh in on this and if we get consensus here then I will add it to my little WiKi (below in my sig) as well.

 

-----signature-----
Windwalker
DAoC Research Wiki: http://tinyurl.com/35564tf
Minstrel Mentor Column: http://tinyurl.com/2uog2ur
Vanesyra  1 star
Posts: 108
Registered: 2009-12-8 17:31:44
oh lol yes it is "groWth". hell i would say i was drunken but i wasn't ^^

@DoorknobMLF
as you said, you need math and statistics knowledge and a very long time on pendragon to figure this out. one example of what i have done there: a level 50 paladina, 1 crush spec. going to a mob write down the unstyled cap damage on a lvl 1 mob, and write down the style damage of the lvl 1 style. go to trainer, train the next style (lvl 2 one). back to the mob, do all styles (lvl 1 + lvl 2 style) again and write down the values of them, and train the next style. and this until you reach lvl 50 spec with all styles. you get a huge set of data. then you use + skill items to get the values for 51 to 61 spec. here is one example picture done from that data 4.4 weapon with 60 quic (picture showing the style damage without your unstyled damage)



only the way how this lines go parallel to each other shows you that growthrate alone can not do this. because you got y = m * x + t (where m = "GROWTH" and t = my "BASE" value).
DoorknobMLF  3 stars
Posts: 627
Registered: 2008-3-2 09:16:57
See even that is a ton of work, but its nice to have. But even then you are testing for the cap damage, which should be the same thing everytime.

I'd be more interested in finding out your expected damage based on a style's growth rate, and your level vs enemies level and everything else involved. In fact it may not be too hard. Start with basic stats, no plus to skills or anything, and find the distribution of unstyled base damage within a certain minimum and maximum range. I would kind of assume that its uniformly distributed (ie. equal chance of hitting for each value between the minimum and the maximim), but it might not be. Can anyone confirm this? If not I might go out and try it myself, it actually wouldn't take too long to test for this.

 

-----signature-----
GardianAngel  2 stars
Posts: 333
Registered: 2003-7-29 21:47:48
Non math guru, looking at this chart, played a crush/shield arms to RR5.


Looking at your chart, I see the "best growth" from the 34-44 styles and

increased dps on all styles at 51-61 for every RR+1. Is this right?


It also looks as if most styles from 10-29 have the same or close growth?

 

-----signature-----
Gardian Angel 50 Paladin
Gardina 50 Warrior
K 73% A 60% S 47% E 27% Ywain?
Hammer 50 Thane
Louden 50 Skald
DoorknobMLF  3 stars
Posts: 627
Registered: 2008-3-2 09:16:57
UnscrupulousDulu posted:

DoorknobMLF posted:

If I played melee classes a lot, and had a lot of time, I would do something like this too, since I have the math and statistics knowledge to do these tests. But there could be a lot of factors acting in many unknown ways, so it could take a very long time to get it right.

These formula's might approximate damage quite well, but i'm no convinced that they are whats actually used in the ingame calculations.

Btw I'm curious, where did you get this info?



"If I played melee classes a lot, and had a lot of time, I would do something like this too,"


First I was all like:


"but i'm no[sic] convinced that they are whats actually used in the ingame calculations."

I'm not skeptical about the way a +10% damage is worked into the formula for damage, this is obvious, same goes with style damage, celerity, and all the other % based bonuses, then its just a matter of finding out how they are applied and this is not hard. I'm talking more about how things like stats and spec level affect damage, weapon speed, etc. While I'm sure the formula's seem to provide the answer, and are at least very close to the answer, i'm not convinced this is exactly how its calculated in the game.

 

-----signature-----

VaultNetwork.net is an independently operated community forum and is not affiliated with, endorsed by, or technically based on IGN, GameSpy, FilePlanet, GameStats, or the former IGN/GameSpy Vault Network.
References to VaultNetwork.net mean this site/domain. VNBoards-style presentation is a visual homage only. By using this site, you agree to the forum rules.