CALCULATING DAMAGE, help to make a correct formula!

[tomiotar]

mmm

Working with this I find a expresion that works really good for me:

Phys = (base + skill_pow * mastery_incr - Phys def) * balance * skill_mult * buff&passive * multiplier

and

Mag=((base + imbue_pow )* mastery_incr - Mag def) * balance * skill_mult * buff&passive * multiplier

The main problem is again the multiplier, because I believe that this number depend on the diference of level between the monster and char. I've found a expresion that work really good for me, but I believe that it will be different at each level (so the formula become useless) because looking at monsters that are 20 level higher level than me (Roc, the unique) my dmg become negative (I'm curing him with my arrows ). The thing is that damage start to be negative, not due to the high defense of the monster but because this "multiplier" factor become negative.

I'm sure that the rest of the formula is correct (including the location of monster defence), because I try different skills with different mobs and it always calculate the correct result, the problem is that I'm sure that as soon as I get lvl86 the formula will not longer work even for me .

Until now I was making this factor just a function of the monster level, now I have to make it a function of the "diference of level" between the char and the monster and find a new expresion .

[NuclearSilo]

Try to test dmg of mob from lvl 1 to 90 to see how the multiplier change. U can skip each 10 lvl mob. I guess that the multiplier will go down, starting from 1.29 (mangyang)

[Nixie]

Everything good and fine, but does anyone know how the damage increase bonuses go into the formula? Surely, you have to multiply the whole formula with the total, but how to get the total? You can just sum them together (like 10% + 20% + 30% = 60%) or multiply them, the right way (1.1 * 1.2 * 1.3 = 1.716 => 71.6%). But I guess that's not important, yet. First we/you have to make to formula work under normal conditions.

[tomiotar]

Try to test dmg of mob from lvl 1 to 90 to see how the multiplier change. U can skip each 10 lvl mob. I guess that the multiplier will go down, starting from 1.29 (mangyang)

Thats that I did yesterday, I even use different multiplier for mag and phys part of the damage (they are close each others but a bit different):





I have to find a formula whith that shape, that depend on the difference of level between char and monsters and also have to be asyntothic to 0, easy job . Using the polynomial function that is porvided on the fitting I get really good prediction for my damage, but the problem is that I'm sure that works only for me or for any other that is lvl 85 like me, but it will be different by any other non-lvl 85.


Just to show how precisely I can calculate the damage using the formula that I'm using but with the correct value of multiplier:

Phys = (base + skill_pow * mastery_incr - Phys def) * balance * skill_mult * buff&passive * multiplier

Non imbue

Devil arrow (lvl81):
Phys= (3491,764 + 590*1,84 - 7 ) * 0,86594 * 3,5 * 1,23 * 1,276772606 = 21753,3



Strong bow (lvl82)
Phys= (3491,764 + 608*1,84 - 7 ) * 0,86594 * 3,5 * 1,23 * 1,276772606 = 21910,9




With imbue:

Total dmg= 21910,9 + ((base + imbue_pow )* mastery_incr - Mag def) * balance * skill_mult * buff&passive * multiplier

Fire:
total dmg= 21910,9 + ((3030,644 + 913)*1,84 - 10) * 0,79348 * 3,5 * 1,20 * 1,287004542 = 21910,9 + 31080,1 = 52991



Light:
total dmg= 21910,9 + ((3030,644 + 45)*1,84 - 10) * 0,79348 * 3,5 * 1,20 * 1,287004542 = 21910,9 + 24229,9 = 46140,8




So, I change phys skills and mag skills and using same multipliers I got a precision on the decimal number to calculate the damage. Most interesting part is the light result when you can see that I'm using the weapon mastery level instead of imbue mastery level to predict the value, so mastery imbue level have no influence on final damage (Actually, I knew this by some experiment that I made at 80 cap when we didn't have 70-80 skills).

[Nixie]

Your formula is preety good. ^^

[bryanvan]

I'll test the formule on my char once im lvl 52

[Nixie]

Uhmm... no, it's not working, lol.
I changed the defense to 7 (mangyang's defense) and it came up with 14k. I've hit 17k some levels earlier if I remember well. ;p

[NuclearSilo]

Uhmm... no, it's not working, lol.
I changed the defense to 7 (mangyang's defense) and it came up with 14k. I've hit 17k some levels earlier if I remember well. ;p

Euro dmg could work differently. But as he said before, the formula works only with his char: it depends on the level or smt else. But it's also possible that the multiplier is universal. Coz all the result i calculated give ~1.29 multiplier. Tomiotar has considered that the dmg and the balance are not rounded and had ~1.27.. as result. And the right result without rounding rule is 1.2856.. smt so i rounded it to 1.29.

so mastery imbue level have no influence on final damage (Actually, I knew this by some experiment that I made at 80 cap when we didn't have 70-80 skills)

From my own experience, it does. Someone use an imbue book 1 lvl 9 but with mastery fire, for ex, lvl 90 will always deal more (much more) dmg than someone with only 30 lvl mastery fire. (with pure str, imbue is too weak to notice)


To make it simple, let's consider the graph of the multiplier is straight line. Because working with the randomness and probability, it's very hard to get the maximum of dmg, except at low lvl. For exemple, rand() * rand(), so the probability to have 0.99999.. * 0.99999.. is 1/10000... . So if u calculate this expression 10000 times, ur not sure to have at least 1 result expected, or even close to it. But results will concentrate at the middle, like a mountain.

lvl 1-30, results concentrate at 80%-100%
lvl 30-60, results concentrate at 50%-90%
lvl 60-90, results concentrate at 20% 80%

Maybe that's why there are 3 curves in the graph. Mob at lvl 90 has high parry ratio, your dmg tends to lean at the minimum dmg side. So try to make another graph in some conditions:
-lvl 1-30: take the low power range
-lvl 30-60: take the middle power range
-lvl 60-90: take the high power range

In your case:
Strong bow craft lvl4 (lvl82 skill)
450 608 (350%)
high = 608
low = 450
middle = (608+450)/2
and also the power in the character info window

This is a way to balance and to have all the result around ~80% of max dmg theoric.

[Nixie]

Uhmm... no, it's not working, lol.
I changed the defense to 7 (mangyang's defense) and it came up with 14k. I've hit 17k some levels earlier if I remember well. ;p

Euro dmg could work differently. But as he said before, the formula works only with his char: it depends on the level or smt else. But it's also possible that the multiplier is universal. Coz all the result i calculated give ~1.29 multiplier. Tomiotar has considered that the dmg and the balance are not rounded and had ~1.27.. as result. And the right result without rounding rule is 1.2856.. smt so i rounded it to 1.29.

so mastery imbue level have no influence on final damage (Actually, I knew this by some experiment that I made at 80 cap when we didn't have 70-80 skills)

From my own experience, it does. Someone use an imbue book 1 lvl 9 but with mastery fire, for ex, lvl 90 will always deal more (much more) dmg than someone with only 30 lvl mastery fire. (with pure str, imbue is too weak to notice)


To make it simple, let's consider the graph of the multiplier is straight line. Because working with the randomness and probability, it's very hard to get the maximum of dmg, except at low lvl. For exemple, rand() * rand(), so the probability to have 0.99999.. * 0.99999.. is 1/10000. So if u calculate this expression 10000 times, ur not sure to have at least 1 result expected, or even close to it. But results will concentrate at the middle, like a mountain.

lvl 1-30, results concentrate at 80%-100%
lvl 30-60, results concentrate at 50%-90%
lvl 60-90, results concentrate at 20% 80%

Maybe that's why there are 3 curves in the graph. Mob at lvl 90 has high parry ratio, your dmg tends to lean at the minimum dmg side. So try to make another graph in some conditions:
-lvl 1-30: take the low power range
-lvl 30-60: take the middle power range
-lvl 60-90: take the high power range

In your case:
Strong bow craft lvl4 (lvl82 skill)
450 608 (350%)
high = 608
low = 450
middle = (608+450)/2
and also the power in the character info window

This is a way to balance and to have all the result around ~80% of max dmg theoric.

It's the same for all players, formula is the same. Europeans just use 1 damage (phy or mag), while Chinese can use both at once. I'll take a look at the formula later and experiment a bit... ^^

[ATOD]

I don't feel like reading through all 5 pages... so who has an accurate damage formula that is universal and works for all level?

[NuclearSilo]

I don't feel like reading through all 5 pages... so who has an accurate damage formula that is universal and works for all level?

Death2U

[bryanvan]

Mind posting it ?

[Nixie]

He's banned. And there is no prove he actually has it, he never published it.

[HolyPure]

[size=75]I am working at the moment on a formula to calculate the damage. I have at the moment a formula that seems to work nice. I guess its not to 100% correct, but i want to test that with more different characters.
[]....bla bla...[]
mage. So use always the highest values and ingore the lower values at first. If you think you have a good formula, then you can calculate the damage again with the lower values to be able to say the damage range of your character.[/color][/b]


greetz, Payne

i wonder why u still worry that 4, must be boring 2 watch ya bot lol (or how many clients u run^^)

B.O.B.S

[NuclearSilo]

i wonder why u still worry that 4, must be boring 2 watch ya bot lol (or how many clients u run^^)

B.O.B.S

Dont turn a good thread into a locked one

[tomiotar]

I've continue working on the formula but I've noticed that I should level up before to see difference between different levels, so:

Phys = (base + skill_pow * mastery_incr - Phys def) * balance * skill_mult * buff&passive * multiplier

and

Mag=((base + imbue_pow )* mastery_incr - Mag def) * balance * skill_mult * buff&passive * multiplier

I will use this values of multiplier:

Phys Multiplier: 1,276772606
Mag Multipier: 1,287004542

First before lvl up (now with 85 pacheon, 85 fire, lvl11 fire imbue. same weapon skills):

Strong bow (lvl82)
Phys= (3510,741 + 608*1,85 - 7 ) * 0,86594 * 3,5 * 1,23 * 1,276772606 = 22030.2




With imbue:

Total dmg= 22030.2 + ((base + imbue_pow )* mastery_incr - Mag def) * balance * skill_mult * buff&passive * multiplier

Fire:
total dmg= 22030,2 + ((3030,644 + 971)*1,85 - 10) * 0,79348 * 3,5 * 1,20 * 1,287004542 = 22030,2 + 31709,4 = 53739,6




I put on red the values that change due to the increase on mastery level and imbue. As you can see the formula still predict the dmg for my char even after using 3 diferent imbues (1 light and 2 fire lvl), changing weapon skills (strong bow and devil arrow) and changing mastery lvl.


Then (finally after 1 month at lvl 85 ^^) I finally level up (this is the reason because you didn't have news from me in last 10 days). So now these are the values that I got after leveling:

Changes lvl85 --> lvl 86:

Stats points
294 --> 296 str
292 --> 295 int

Balances
str balance= 86,594 --> 86,380
mag balance= 79,348 --> 79,301

Attacks
base phys attack= 3510,741 --> 3518,389
base mag attack= 3030,644 --> 3040,565

Aplying the formula again:

Strong bow (lvl82)
Phys= (3518,389 + 608*1,85 - 7 ) * 0,86380 * 3,5 * 1,23 * 1,276772606 = 22012,1



Fire:
total dmg= 22012,2 + ((3040,565 + 971)*1,85 - 10) * 0,79301 * 3,5 * 1,20 * 1,287004542 = 22012,1 + 31769,3 = 53781,4




So I'm 1 level higher but the formula continue predicting the results without a change on the multipliers. I don't know if still work the formula because the changes are too small to notice them (I'm just 1 lvl higher) or because these multiplier values don't change with lvl (something that until today I was sure it was true). I need another chinesse char but with a higher diference of level between me and him to check the formula. I continue doing it wih mangyan because we are looking for a general formula that can be applied for any build (first I want to see chinesse chars and then look for correct multiplier for euor chars), and mangyans are the monsters that everyone at any level can kill without problem ^^. The chinesse char have to use weapon skills (not nukes) and post a screen of his weapon, and tell us the masteries lvl and skills he is using. Help will be apreciated .

An interesting point is that formula predicts the decrease on phys dmg, because of the decrease on the phys balance is more important than the increase on base phys attack. In the case of mag balance the decrease is much lower so mag dmg go higher due to the extra int points when I level.

Now, look that you make me do to continue with the post =P:

[Nixie]

It's good to see it's still working. Going to test it on my warrior, if it will work, noticed some mistakes in my program which are solved now.

Btw... all the formulas lack 2 things. Attacker's attack ratio and the defender's parry ratio (monster parry). I'll try out a few think to see how the whole parry ratio works and I'll try to insert them into the formula. Hope to com back with good results. =)

[tomiotar]

It's good to see it's still working. Going to test it on my warrior, if it will work, noticed some mistakes in my program which are solved now.

Btw... all the formulas lack 2 things. Attacker's attack ratio and the defender's parry ratio (monster parry). I'll try out a few think to see how the whole parry ratio works and I'll try to insert them into the formula. Hope to com back with good results. =)

Parry and attack rating don't change the maximum value that you can get, they change how often you get that dmg, thats the reason because I use always highest possible dmg (I hit lots of mobs until I see that the higher dmg repeat excactly same number) instead of intermediates values. I have pacheon passive + blue hawk + bow with 71% on attack rating white stat, so I found much more frequently the higher dmg than the lower one (I can use lower value too, but I rarely see a low value that repeat).

EDIT: I'm not sure if euro use same multipliers than chinesse. If you find that dmg doesn't fit for mangyang change multiplier to make it fit and then try with a different weapon skill, using same multiplier that you just find.

[Nixie]

It's good to see it's still working. Going to test it on my warrior, if it will work, noticed some mistakes in my program which are solved now.

Btw... all the formulas lack 2 things. Attacker's attack ratio and the defender's parry ratio (monster parry). I'll try out a few think to see how the whole parry ratio works and I'll try to insert them into the formula. Hope to com back with good results. =)

Parry and attack rating don't change the maximum value that you can get, they change how often you get that dmg, thats the reason because I use always highest possible dmg (I hit lots of mobs until I see that the higher dmg repeat excactly same number) instead of intermediates values. I have pacheon passive + blue hawk + bow with 71% on attack rating white stat, so I found much more frequently the higher dmg than the lower one (I can use lower value too, but I rarely see a low value that repeat).

EDIT: I'm not sure if euro use same multipliers than chinesse. If you find that dmg doesn't fit for mangyang change multiplier to make it fit and then try with a different weapon skill, using same multiplier that you just find.

You may be right about attack/parry rating. I was told that if you can deal 300~500 damage and have higher AR you'll hit for example 400~500, if you'll have low AR you'll hit 300~400. Never didn't really know how exactly it works.

And yes, I'll try it later if I'll be able to connect to the server. ;_;
My multiplier is probably at about 1,34 or how much it was the last time (but I had a little mistake in my program, used skill% instead of skill attack... <.<).

[Nixie]

I've been just testing out something, and the Silkroad percent increase will always sum up first and then increase the original value. The best is I show you an example...
This is how Silkroad does it:
x = basic_value * (1.18 + 1.75)

This is how it should be done actually (at least I think so):
x = basic_value * 1.18 * 1.75

So in future, when applying bonuses anywhere, just sum the percentage together and multiply the base with it. It might be old, but those who don't know it yet (I didn't really know which method they used to sum them together ^^') should find it helpfully I guess. ;p

[HolyPure]

i wonder why u still worry that 4, must be boring 2 watch ya bot lol (or how many clients u run^^)

B.O.B.S

Dont turn a good thread into a locked one

oi lol u dont like the truth ? lol i shit on the topic

[Stress]

[OFF-TOPIC]
Tomiotar: 90 fire / 90 ice / 90 pach / 30 light?

[Nixie]

[OFF-TOPIC]
Tomiotar: 90 fire / 90 ice / 90 pach / 30 light?

Lightning is more than 30, look at his speed buff. ^.-

[tomiotar]

At lvl 90: 90 pacheon/fire 60 light/cold
At this moment: 85 pacheon/fire 70 cold 56 light (I need those 4 levels of light now to make 0 gap until lvl90, then farm at 5 gap, then delevel ice to 60).

[IceCreamMan]

i dont think brute forcing a formula is gonna work. you can't brute force an equation and then find a multiple to add to it to get the correct dmg. it wont work for other levels. just a suggestion

[tomiotar]

As it seem that anyone else is trying the formula I decide to use a char that I made some months ago because I want to see the actual difference between crits on pur str, hybs and pure int. The idea it was make 3 lvl24 chars, using npc gear and npc lvl24 bow with fire imbue and antidevil+ phys and mag buff. I made the pure str and then I got bored so thats the reason because I didn't post the results. So I decide to use this char with the system that I use with my main char. Test data

Char lvl 24:

112 str
43 int

Phys balance: 93,548
Mag balance: 34,677

Weapon NPC bow lvl 24
Phys attack: 110
Mag attack: 176
Phys rein: 59,0
Mag reinf: 94,4

Skill
Andidevil missil lvl9 42 - 57 (150%)
Imbue
River Fire force lvl9 50-83

Buff: 5% Phys dmg (3% buff + 2% passive) and 5% mag dmg


So I apply the formula

Phys = (base + skill_pow * mastery_incr - Phys def) * balance * skill_mult * buff&passive * multiplier = (218,34 + 57*1,24 - 7) * 0,93548 * 1,5 * 1,05 * 1,276772606 = 530,5




Total dmg= 530,5 + ((base + imbue_pow )* mastery_incr - Mag def) * balance * skill_mult * buff * multiplier = 530,5 + ((216.59 + 83)*1,24 - 10) * 0,34677 * 1,5 * 1,05 * 1,287004542 = 530,5 + 254,1 = 784,6



So, even when I didn't get same value, the formula was really accurate using another skill, another build (pure str against hyb 1:1) and 61 lvls difference. I want more people with other builds (not only pacheon) to try the formula, kill lvl 1 monster isn't that hard, the only boring thing is that you have to shoot them several times to be sure that you are getting max dmg (when you see that high dmg you get is repeat you can stop). Just remember the requisites: chinesse, weapon skills (not nuke) and if it is possible with and without imbue. If you don't know the formulas to calculate balances and base attack just ask for them or tell me the amount of str and int and the data of your weapon (attacks and reinfs) and I will calculate them for you.

I will organize all the multipliers that I've found and post it later (I have multiplier values for monsters lvl 1, 10, 20, 30, 40, 50, 61, 70, 80 and 88) and the polynomial function that I've found to correlate them (using the function you can apply the formula to all the monsters instead of just mangyans, but is a bit less accurated).


BTW

i dont think brute forcing a formula is gonna work. you can't brute force an equation and then find a multiple to add to it to get the correct dmg. it wont work for other levels. just a suggestion

Thanxs for the advice, but I don't consider that changing 1 by 1 each possible variable and then measuring the effects can be consider brute forcing. The experimental approach is one of the most used systems on science, when you need and answer and you don't have the theoretical model to descrive an event .

[selenne]

[off-topic] im lovin' to read this thread...

[chesticles]

[off-topic] im lovin' to read this thread...

All your posts are off topic lol

OT: I'm using the information in this thread to make a damage calculator using VB.

ADD: I see Nixie is already working on one. Mine will be prettier xD

[NuclearSilo]

Yo.
Good job tomiotar. We have advanced one more step in the research of the dmg formula.

After we find a close multiplier, it's possible to use the formula above to find which build deal the best dmg crit+imbue to Mangyang.

From your tests above, i have some deduction, that the multiplier is:
- independant from the character's level
- dependant on the mob level (or type of mob)
- dependant on the character's int/str distribution

The 3rd deduction, i noticed long time ago. If u check the result with my old formula i posted in page 3-4, u'll see that the 1.29 multiplier work with mangyang for pure Str and hybrid, and not pure int which give a big error.

Hypothesis: Is increasing INT increase the magical multiplier (same with STR)?
Doesnt the multiplier act like the balance? (one go up, one go down...)

Try also to find if it's writen anywhere in the PK2 (characterdata.txt) a value close to the multiplier, it could be a base multiplier.

[tomiotar]

This is the list of multipliers that I've used to find these graph:


Level Phys multiplier Mag Multiplier Monster
1 1,276772606 1,287004542 Mangyan
10 1,171651277 1,181220622 Decayed yeowa
20 1,071583454 1,080130069 Chakjy
30 0,984744926 0,991719079 Earth ghost magician
30 0,983911534 0,992490493 Hyengocheon
40 0,905604863 0,912703332 Red Scorpion
50 0,831677002 0,837808077 Ujigi
61 0,751189613 0,756244326 Soil Ghost bug
61 0,749117844 0,762497297 Shakram
70 0,678674164 0,688549128 Cloak Mask
80 0,521141074 0,531627141 Wing Tribe
88 0,391522552 0,391718496 Rocky

Acording to this correlation the formula to calculate multipliers as a function of the level is:

Phys Mult: 1,29731 - 0,01497 * X + 0,0002 * X^2 - 1,6767*10^(-6) * X^3

Mag Mult: 1,30921 - 0,01543 * X + 0,00021 * X^2 - 1,7788*10^(-6) * X^3

(X= monster level)

Take into account that multiplier values calculated by the formula are just aproximations but they will lead to error (you will not get the same value but they will be close). Another interesting point is that the monsters of same level that have higher phys defence than another of same level will have lower phys multiplier (so, he recieve less phys dmg), but it has higher mag multiplier (so, he recieve more mag dmg). I'm not sure about the meaning of this, because maybe the monsters have "balances" to affect the dmg or maybe the mulltiplier is a function too of (phys def)/(mag def) ratio.