Bright's Spreading Flame Rune damage formula?

[Punkaiser]

I tested lv42 Bright hit on 2 Copper Suns:
381, Spreading Flame 301
after that battle he leveled to 47, still has only 1 swing, hit on 2 Copper Suns again:
411, SF 328
All the cases tend to show 0.8x multiplier, i guess it's safe to conclude now.

[KFCrispy]

what's the PDF of Copper Sun?
the problem with all of your calculations is that you are comparing damage, which is ATK - DEF. most special attack multipliers are applied to the ATTACK part before defense is applied, not DAMAGE. fight a different enemy with significantly more or less defense, and most likely you'll get significantly different ratios between the final damage numbers.


I strongly believe we should try to solve this equation instead (without the complications of the Damage skill or damage/atk boosting runes):
(ATK + x)*y - DEF = DMG
x is the possible increased ATK bonus (I'm going to assume 100 for Spreading Flame but you can try 0)
y is the assumed ATK multiplier

For multiple swings it becomes more complicated but it's the same 2 variables.
Swing 1 = (ATK + x)*y - DEF
Swing 2 = (ATK + x)*y*0.75 - DEF
Swing 3+ = (ATK + x)*y*0.75*0.75 - DEF
DMG = Swing 1 + Swing 2 + ...

Many people make this mistake, which is why they often come up with 0.7 multipliers for successive swings--they just fight 1 enemy and get numbers from that but fighting an enemy with different PDF will give them wildly different damage ratios.


with my proposed formula (+100 base dmg and 2/3 multiplier to ATK) and using on your numbers i get 50 for the Copper Sun's PDF, but my numbers only work if you don't have the Damage skill or any other multipliers active... would mean Bright's original ATK was about 431 and then 461 after leveling up.

[Punkaiser]

All the numbers i gave you were damage output which are already after math, they are not attack power showed on status screen.
By "381, Spreading Flame 301" i meant bright direct attack Copper sun for 381 and causes 301 of damage with SF. We already had 4 samples of tests from 1 swing to 4 swings with all results are aftermath, is there really a need to dig to the root of calculation?

And i find there's problem with formulas above.
Suppose x=0 and y=1.15 (since i always test bright with B+ damage skill)also suppose his attack value=400, foe's def=100
the first swing with your formula 'd be=400.1.15 - 100 = 360
while the 2nd swing with that formula= 400.1.15*0.75 - 100 = 245
now compare 2nd and 1st = 245/360 *100%= 68.06% which is wrong as i see the 2nd swing always deal about a perfect 0.7x or 0.75x of the 1st and it even goes worse if we remove 1.15 in those calculations. That's why i think the formula should be (ATK-DEF+x)*y
it may sound too simple but seems to make more sense with result in test like 473 + 331 + 231 + 231, trust me, i wasnt rounding any number and this happenned thrice in a row.

[KFCrispy]

The Damage skill boosts damage, so the multiplier goes outside of the ATK-DEF piece, just like Double-Strike's multiplier.
ATK modifiers like successive swings are applied to the ATK piece, and I'm claiming most or all special attacks actually affect ATK not DMG. However, I haven't tested these minute details because they definitely haven't been a concern.. but I'd like to at least get the mechanics behind Spreading Flame and Devil's Doll. don't worry, I will probably confirm your formulas or mine in less than a month

it's worthy to note that the berserk status affects ATK or DMG depending on the game. In Suikoden II we had confirmed it affects ATK (people got better damage results from berserk than from Double-Strike's 1.5x damage modifier in that game), while in Suikoden V we had confirmed it affects DMG (which is why Power Rune's 1.5x ATK gives you better results).

[eldrasidar]

so you are saying damage goes something like this:
ATK - Monster Defense = damage dealt on first swing x (damage skill bonus, double strike bonus, song of madness bonus, possibly fury bonus), then multiplied by .7-.75 for a second swing, and so on for each additional swing.

[KFCrispy]

no the swing ratio is applied directly to ATK before enemy PDF reduces it. but the Damage skill and Double-Strike definitely affect the (atk-def) part.
Damage of a single attack = {(ATK + x)*y - DEF}*z
ATK is your base ATK value
x is the rare Devil's Doll / Spreading Flame that I proposed (of course, I could be wrong)
y is the set of ATK modifiers (i expect special attack modifiers are placed here)
z is the set of damage modifiers

For special attacks you need to question if the total ATK from multiple swings (each swing still having a hit/miss ratio) actually add up into ATK, or if each swing is applied against defense.

[Punkaiser]

I undertand your unease with that S5 formula since stuff like power rune kick in ATT value. But look at the flaw in your formula if we try to calculate damage difference between the swings
I could re-take the last exemple of ATT=400, foe DEF=100, let's ignore other factors
1st swing= 400 - 100=300
2nd swing= 400*0.75 - 100=200
=> 200/300=66% and if enemy DEF rises higher to 200 for exemple then 2nd swing damage would equal to only 50% of 1st swing DMG
so i think its not right to put that 0.7-0.75 multiplier in as "y" but as "z" is more accurate which means it affects DMG value, not ATT value

[KFCrispy]

yes some people recorded 2nd swing ratios of 40-50% at times. the way i confirmed the multiplier is definitely higher than 70% is through observing critical hits which completely ignore enemy defense. the critical hit ratios were consistently between the 1/2/3 swings is 73-76% which normal hits that use enemy defense cause it to look more like 65-70% (and sometimes 30-40% for low ATK characters). that was my proof for why it affects ATK.

[Punkaiser]

Then i suggest you to really try out Mel , lol
you will notice that during her 6 swings, very often from the 3rd to 6th which are meant to be equal, you will get very different output while no "guard" message pops up at all, for this reason shes the unstable queen. If we only observe her direct hits for a few times we may fall in that 30-40% conclusion while there are times she'd cause a full 75% non critical. I'd suppose there were alot of hits werent able to overcome enemy defence. For the time being, i think your calculation for the 2nd = 0.7-0.75 of the 1st is exact and quite absolute, but we're fooled around by inconsistent damage from observing chars like Mel, i boosted her attack value to 400 and skill to 200 but didnt help so the theory that it happens exclusively to chars with low ATT failed.

I don't think i ve encountered anything weird from bright performance regarless the enemy i fought. I've checked their Skill growth, your source is still correct, Mel's skill is same as Bright's, then now its questionning me why exists such unstable performance for the girl?

[KFCrispy]

i just set Bright's str to 255, making his ATK 637. he hit extremely consistently as Punkaiser observed--- got the exact same number 2-3 times. (so what's up with Mel???) i used the weaklings in Brass Castle, focusing on the Dokugumo
normal: 612
spreading flame: 734
perhaps this enemy is weak vs fire? i'll hit something else...

[Punkaiser]

i just set Bright's str to 255, making his ATK 637. he hit extremely consistently as Punkaiser observed--- got the exact same number 2-3 times. (so what's up with Mel???) i used the weaklings in Brass Castle, focusing on the Dokugumo
normal: 612
spreading flame: 734
perhaps this enemy is weak vs fire? i'll hit something else...

suppose spreading flame deal 0.8x => technically damage= 0.8 * 612= 489.6
in case the enemy has a weakness againts fire which make it suffer x1.5 damage from fire=> SF= 1.5*489.6= 734.4
It seems exactly be your case.

so your ps2 is up?

[KFCrispy]

yeah my ps2 has been up but i've been busy. just putting a little free time into this right now...

note: also lowered Bright's speed so he only hits once

i'm going to ignore Dokugumo because the fire weakness is troublesome - not sure if the enemy takes double dmg or something like 75% or 50% more.
tested 2 more enemies, the Troll Dragon in Amur Plains and Songsprite in Budehuc Basement.
vs Troll Dragon
normal: 577 (enemy def = 60)
spreading flame: 403

vs Songsprite
normal: 622 (enemy def = 15)
spreading flame: 435

vs Copper Sun
normal: 612 (enemy def = 25)
spreading flame: 489

so..... i can't come up with a formula! maybe MDF/elemental resistance is involved?
i got the same 80% damage ratio as Punkaiser vs Copper Sun which is definitely indicating *something*, but vs other enemies with more and less PDF, the ratio was 70% for both cases.

since the damage ratios were about the same when Bright's ATK was really high vs a 'normal' range on the same enemy, it's suggesting the +100 fire damage in the description is not being used.

next i'll try lowering his STR to really low number.

[KFCrispy]

ok with a really low ATK (str=20 --> ATK=50)
i got the same ratios. vs Song Sprite, 70%. i also hit the Svire in Budehuc's basement, 80%

so the "100" in the description is definitely not "+100 damage". the attack uses a ratio based on the normal attack damage. to explain the 70% vs 80% and that Dokugumo 120% damage, we have to assume elemental weakness is definitely factored in. is the normal ratio 70% or 80%? we'd have to figure out which of those enemies are slightly weaker vs Fire (or general magic spells) vs which take normal damage.


this is a far-fetched thought, but what if it actually deals 100% damage (the possible 100 from description) but the enemies we've been hitting all have some form of elemental or general magic resistance (remember there's that Magic Resistance skill)?

[Punkaiser]

I'm pretty positive that fire element does kick in if you still remmember my post somewhere about Mel and Bright alway deal an additional modifier x1.5 with their rune on twin snake body or x2 on the old granny tree in kuput forest.

[Punkaiser]

Too bad i've formatted my mem card, can't test it right now but i i guess we gotta determine enemy fire resistance in order to go further :

Steps to take:
-delete fire skill to set fire spell damage to 100%
-use dancing flame for exemple on test objects
-Do a math A= user's magic minus enemies Mdef plus 150 from spell
-B= actual damage shown on screen, then enemy fire affinity should be = A/B

Once we could find out which monster is neutral againts fire then our test for SF would go more precisely.