Attack Level and Defense Level Demystified
Comments
-
Legerity - Sanctuary wrote: »I really recommend you read the thread.
+100atk vs +0def level isn't double damage
+100def vs +0atk isn't half damage or what you're implying either
+1atk cancels +1def
It appears to be a function comprised of multiple stepwise functions. Go read the thread to find at what the threshold is where the function is no longer linear.
Yes; I proposed a linear funtion (Ax+B) for atk lv > def lv , and a inverse linear funtion C/(x-D) for def lv > atk lv. Thats right.
I know that +1 atk cancels +1def. The formula I gived dont break that rule, despite the defense is overweight.
I ve only glanced the thread, and read some parts. Ok I ll see all carefully.
But can you tell me since now, in what part of the thread there are experimantal datas about the no linearity, WHEN ATK LV IS BIGGER THAN DEF LV.
And in what part of the thread are there experimental information about the damage of a +100 atk level over a +0 def level. I did a fast searching and I didnt find this information.0 -
Yes; I proposed a linear funtion (Ax+B) for atk lv < def lv , and a inverse linear funtion C/(x-D) for def lv < atk lv. Thats right.
I know that +1 atk cancels +1def. The formula I gived dont break that rule, despite the defense is overweight.
I ve only glanced the thread, and read some parts. Ok I ll see all carefully.
But can you tell me since now, in what part of the thread there are experimantal datas about the no linearity, WHEN ATK LV IS BIGGER THAN DEF LV.
And in what part of the thread are there experimental information about the damage of a +100 atk level over a +0 def level. I did a fast searching and I didnt find this information.
This is where it gets messy. My original discomfort of you posting was because of:
1. You proposed a generic methodology that these testers know. I recommend you dig through some Archer research in the useful link section because it shows that the people who run these test on PW mechanics do know how to construct experiments.
2. Your second post basically reiterates that +30atk level = +30% damage which was already done.
The actual threshold is still unknown but there's a general indication of where it lies based on here.
A basic summary can be found here. And of course, most of what has been discussed can be found on the first post.
When I told you to "read the thread" I was actually holding a gripe on your +30atk lvl = +30% damage because it really seemed like you didn't read the thread.
Oh and my mistake, the atk lvl statement I made above (and previously) is false. It's linear when +def lvl is not involved.
edit: fixing some linking issues - browser also going haywire on me...0 -
Legerity - Sanctuary wrote: »This is where it gets messy. My original discomfort of you posting was because of:
1. You proposed a generic methodology that these testers know. I recommend you dig through some Archer research in the useful link section because it shows that the people who run these test on PW mechanics do know how to construct experiments.
2. Your second post basically reiterates that +30atk level = +30% damage which was already done.
The actual threshold is still unknown but there's a general indication of where it lies based on here.
A basic summary can be found here. And of course, most of what has been discussed can be found on the first post.
When I told you to "read the thread" I was actually holding a gripe on your +30atk lvl = +30% damage because it really seemed like you didn't read the thread.
Oh and my mistake, the atk lvl statement I made above (and previously) is false. It's linear when +def lvl is not involved.
edit: fixing some linking issues - browser also going haywire on me...
I ve read the links you gived me. And here are my conclusions:
1. The formulas:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,2* (def lv - atk lv) / 100 ) )
FIT WELL IN ALL THE TESTS, what I think is the important point.
2. About the comment you said me "The actual threshold is still unknown but there's a general indication of where it lies based on here".
Im not sure of be understanding well what you mean; but I think you mean that it cant be linear because there arent fix ratios here:
"Next, for the DEF values which were tested at both 0 and 20 ATK, I divided the two to find the ratio of normalized damage.
Code:
DEF Ratio
0 1.20
5.5 1.22 (actually 5.5 and 5, which probably accounts for this being high)
36.3 1.20
41.8 1.19
76 1.14"
If it is this, let me tell you that this happen because, we are combining 2 different formulas and the second is not linear for the defense level, neither for the attack level. But this dont deny the linearity of the first formula.0 -
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,2* (def lv - atk lv) / 100 ) )
FIT WELL IN ALL THE TESTS
They do not seem to be completely accurate here? http://pwi-forum.perfectworld.com/showpost.php?p=7777152&postcount=200 -
They do not seem to be completely accurate here? http://pwi-forum.perfectworld.com/showpost.php?p=7777152&postcount=20
Lets see how much is the error.
"+atk lv"
"+def lv"
damage
basis damage
formula prediction
percentaje of error
0
0
1227
1227
1227
0%
0
41,8
822
1227
817
0,6%
0
36,3
856
1227
855
0,1%
20
41,8
980
1227
973
0,7%
20
36,3
1029
1227
1026
0,2%
0
5,5
1157
1227
1151
0,5%
20
5
1411
1227
1411
0%
20
0
1472
1227
1472
0%
The error is always less than 1%. I dont think it is too much, maybe this variation comes of the incertitude of the experiment.
But this theme is still open.
It would fit even better if I had used the formula:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,18* (def lv - atk lv) / 100 ) )
The error would be in this case:
"+atk lv"
"+def lv"
damage
basis damage
formula prediction
percentaje of error
0
0
1227
1227
1227
0%
0
41,8
822
1227
822
0%
0
36,3
856
1227
859
0,3%
20
41,8
980
1227
976
0,4%
20
36,3
1029
1227
1029
0%
0
5,5
1157
1227
1152
0,4%
20
5
1411
1227
1411
0%
20
0
1472
1227
1472
0%
But its difficult to accept for me that the formula contains an arbitrary long decimal number like 1,18.0 -
asterelle - Sanctuary wrote: »i Just Did A Test To Get Some More Data With A High Level Psychic.. Here's The Data From Before
And Here Is The New Data With Even Higher Defense Level Along With Predictions From Proposed Formulas.Att Def Dam (a-d)/100 1.01^(a-d) .99^(d-a) 0 0 836 836 836 836 0 66 466 284 433 430 0 76 437 200 392 389 20 76 500 367 478 476
Defense Level Is Less Effective At High Defense Levels Than Any Proposed Formula. At High Defense Levels Every Defense Level Seems To Further Reduce Damage By Only About 0.6% More. I Don't Yet See A Grand Unified Formula That Works Over The Full Range Of Values. Fiting A Formula To High Defense Levels Doesnt Work Well With Lower Levels And Vice Versa.
Any Guesses? I'd Really Like To Fix The First Post.
edit: Fixed Copy Paste Error
Using these formulas:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,18* (def lv - atk lv) / 100 ) )
We predict the following:Att Def Dam (a-d)/100 1.01^(a-d) .99^(d-a) my prediction 0 0 836 836 836 836 836 0 66 466 284 433 430 470 0 76 437 200 392 389 441 20 76 500 367 478 476 503
And using these other formulas:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,2* (def lv - atk lv) / 100 ) )
We predict the following:Att Def Dam (a-d)/100 1.01^(a-d) .99^(d-a) my prediction 0 0 836 836 836 836 836 0 66 466 284 433 430 467 0 76 437 200 392 389 437 20 76 500 367 478 476 500
0 -
In conlusion it seems that he correct formulas are:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( (atk lv - def lv) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,2* (def lv - atk lv) / 100 ) )
And that the variables like character levels, defense and resistances dont enter to the equation.0 -
Ok, have I understood it right that adding defence lvls up to 40, will add yield an equal return on investment for every +2 Def shard you add?
Where as every +2 Def shard you add after +40 Def lvls will give you diminishing returns?
So with 5 pieces of Nirvana gear, and event neck and belt, you would gain the most return by adding 14 Jades of steady defence, and adding more of these will be less effective, to a point where adding +10Vit or Citrine gems will add you a larger % increase in survivability?
Sorry if I'm mistaken, I'm just trying to translate all this giberish into some info that is useful to my actual gameplay b:cute0 -
Lets see how much is the error.
"+atk lv"
"+def lv"
damage
basis damage
formula prediction
percentaje of error
0
0
1227
1227
1227
0%
0
41,8
822
1227
817
0,6%
0
36,3
856
1227
855
0,1%
20
41,8
980
1227
973
0,7%
20
36,3
1029
1227
1026
0,2%
0
5,5
1157
1227
1151
0,5%
20
5
1411
1227
1411
0%
20
0
1472
1227
1472
0%
The error is always less than 1%. I dont think it is too much, maybe this variation comes of the incertitude of the experiment.
We do not have enough data for enough different cases to say that the error will always be less than 1%.
Meanwhile, 1% can mean the difference between winning and losing.0 -
Archers must be so smart people
my head hurt reading thisb:sad0 -
Templar - Sanctuary wrote: »Archers must be so smart people
my head hurt reading thisb:sad
No, we are not necessarily smart... just... overeducated?0 -
We do not have enough data for enough different cases to say that the error will always be less than 1%.
Meanwhile, 1% can mean the difference between winning and losing.
It is my last proposal, while you dont have a large amount of experimental information or a better formula:
if (atk lv > def lv) then: damage taken = damage delivered * ( 1 + ( {atk lv} - {def lv} ) / 100 ) )
if (atk lv < def lv) then: damage taken = damage delivered / ( 1 + (1,2* ( {def lv} - {atk lv} ) / 100 ) )
where {} means enter part or truncate the number; for example {41,8}=41.
Using this last two formulas we get the following:"+atk lv" "+def lv" basis damage damage damage prediction 0 0 1227 1227 1227 0 41,8 1227 822 822 0 36,3 1227 856 857 20 41,8 1227 980 980 20 36,3 1227 1029 1029 0 5,5 1227 1157 1158 20 5 1227 1411 1411 20 0 1227 1472 1472
Pd. I am a wizard0 -
Balthier - Dreamweaver wrote: »Ok, have I understood it right that adding defence lvls up to 40, will add yield an equal return on investment for every +2 Def shard you add?
Where as every +2 Def shard you add after +40 Def lvls will give you diminishing returns?
So with 5 pieces of Nirvana gear, and event neck and belt, you would gain the most return by adding 14 Jades of steady defence, and adding more of these will be less effective, to a point where adding +10Vit or Citrine gems will add you a larger % increase in survivability?
Sorry if I'm mistaken, I'm just trying to translate all this giberish into some info that is useful to my actual gameplay b:cute
Actually in a first aproximation, the inversion of +2Def lv will give you a equal return on investment no matter how much Def lv you already have; its just the same that happens with the hp, the physical defense or the elemental resistances.0 -
Ive just realized that these formulas fit exactly in all the experimental datas displayed:
if (atk lv > def lv) then:
damage taken = { damage delivered * ( 1 + ({ atk lv } - { def lv }) / 100 ) }
if (atk lv < def lv) then:
damage taken = { damage delivered / ( 1 + (1,2 * ({ def lv } - { atk lv }) / 100 ) ) }
Where {} means enter part or truncate the number; for example { 367,9 } = 3670 -
So at end is it good going full def lvl in armors or no?
b:chuckle0 -
Templar - Sanctuary wrote: »So at end is it good going full def lvl in armors or no?
b:chuckle
I am lv 70, so I dont have experience with the end game gear; neither with the highest levels characters.
An accurate answer its not simple, because it depends on many things; depends on how much you have in almost every state that appear in your character paper (hp, physical defense, elemental reistances, defense level, ...); depends on if you have buffs that increase some states; depends on the level of the person/mob that is attacking to you. Depends on your hp regeneration (this include powders, pots and healing, in case of be making use these things continuously) and depends on the intensity of the damage for second delivered by the person/mob is attacking you; in every distinct case, each state have somewhat different worth. And it depends too on your interests, if you rather be somewhat better in short combats where you recieve ultra high damage for second, or in long combats.
But I can tell you 2 things without more information.
1. Maybe go for full +def wont be the absolutely best possible option for you, or maybe it will be actually the absolutely best possible option for you. But what is always true, is that go for full +def never will be a bad option for anyone with decent hp, defense and resistances. Even more, I believe that full +def could be the best way to go in most of the cases.
2. And a rule that is always true.
If "a" is what you have of +def; "b" is what you have of hp; "c" is the amount in which you can increase your +def and "d" is the amount in which you can increase your hp instead increase your +def.
Always that: "(1+1,2(a+c)/100)/(1+1,2a/100) >= (c+d)/c" you must choose the +def over the hp.
but "(1+1,2(a+c)/100)/(1+1,2a/100) < (c+d)/c" dont necessary mean that you must choose the hp over the +def.
For example in my personal case the last proposals yields that +2def lv worth at least 82hp, but in my personal case regarding all my information and my personal interests; Ive ponderend that a +2 def lv worth for me in my current game somewhat more than 130hp.
And in my particular case +2def is equivalent too, to (+80 of physical defense, +118 of metal resistance, +118 wood resistance, +118 fire resistances, +124 of water resistance and +95 earth reistance)
Note: this information go over the gear; you need multiply this datas for such coeficients in order you can calculate your final states of physical defense and elemental resistances.
My information:
I am a well equipped lv 70 light armor wizard.
vit 29 (this dont matter to the calculation)
str 78 (this dont matter to the calculation)
dex 74 (this dont matter to the calculation)
mag 218 (this dont matter to the calculation)
Basis physical damage: 606-856 (this dont matter to the calculation)
Basis magical damage: 3174-3597 (this dont matter to the calculation)
Chance of critical: 4% (this dont matter to the calculation)
3424 of hp
4253 in physical defense with my own buff
3450 in metal, wood and fire resistance
3777 in water resistance
6240 in earth resistance with my own buff
494 evasion
0 +defense level
0 -%of physical damage taken
0 -%of magical damage taken
4 hp/sec of natural hp recovery
physical defense coeficient, with my str, vit attributes and my own buff: 2,12
metal, water, wood and fire resistance coeficient, with my mag and vit attributes: 1,28
earth reistance coeficient with mag, vit attributes and my own buff: 2,28
And my speciality or main interest:
fight against 3 magical lv70 mobs or 5 melee lv70 mobs at the same time (kill 5 melee mobs is easier for me than kill 3 rang mobs because with my aoe skill dragon breath, they die at the same time pretty fast, and I last almost the same time alife being hitted by 5 melee mobs that being harmed by the magical attacks of 3 magical mobs); all of this using hp powder (+50 hp recovery), using lv60 health pots continuously while I am in that intense combat, and using second wind like reserve.
EDIT: Actually regarding that my second wind give me 1512hp more hp; the +2def lv would worth for me somewhat more than 188hp.0 -
Ive just realized that these formulas fit exactly in all the experimental datas displayed:
if (atk lv > def lv) then:
damage taken = { damage delivered * (1 + ( { atk lv } - { def lv }) / 100 ) ) }
if (atk lv < def lv) then:
damage taken = { damage delivered / ( 1 + (1,2* ({ def lv } - { atk lv }) / 100 ) ) }
Where {} means enter part or truncate the number; for example { 367,9 } = 367
Oooo very nice. I'll try to confirm this with some more data.
EDIT Now with the jones blessing I'm able to test higher attack levels and can confirm the linear formula is exact up to an attack level of 55DEF ATT DAM 1+(A-D)/100 0 0 643 643 0 5 675 675 0 20 772 772 0 25 804 804 0 30 836 836 0 35 868 868 0 50 965 965 0 55 997 997
[SIGPIC][/SIGPIC]
Refining Simulator - aster.ohmydays.net/pw/refiningsimulator.html (don't use IE)
Genie Calculator - aster.ohmydays.net/pw/geniecalculator.html - (don't use IE)
Socket Calculator - aster.ohmydays.net/pw/socketcalculator.html0 -
Asterelle - Sanctuary wrote: »Oooo very nice. I'll try to confirm this with some more data.
EDIT Now with the jones blessing I'm able to test higher attack levels and can confirm the linear formula is exact up to an attack level of 55DEF ATT DAM 1+(A-D)/100 0 0 643 643 0 5 675 675 0 20 772 772 0 25 804 804 0 30 836 836 0 35 868 868 0 50 965 965 0 55 997 997
Excuse me for the question; Im curious to know about it. Where did you get a +lv55 blessing jones?0 -
In conclusion, the correct formulas to calculate the effect of attack and defense level are:
if (atk lv > def lv) then:
damage taken = { damage delivered * ( 1 + ({ atk lv } - { def lv }) / 100 ) }
if (atk lv < def lv) then:
damage taken = { damage delivered / ( 1 + (1,2 * ({ def lv } - { atk lv }) / 100 ) ) }
Where {} means enter part or truncate the number; for example { 367,9 } = 367
NOTE: What I mean with damage deliver is the the damage that would do if the attack and defense level attributes werent exists, but it consider all the other deductions, like the defense and resistances deductions or the difference of characters and mobs level penalizations.0 -
Excuse me for the question; Im curious to know about it. Where did you get a +lv55 blessing jones?
+30 from blessing, +20 from frenzy, +5 from shards/gear?
I'm just assuming. =/Proving that not only archers do math since 2009. b:victory
Current math challenge: pwi-forum.perfectworld.com/showthread.php?t=1029711&page=45
[SIGPIC][/SIGPIC]
"Any skills that can be used to kill you will interrupt BB when successful." -truekossy | "...Sage archers are kind of like Mac owners. They are proud of the weirdest and most unnecessary things." -Aesthor | "We ALL know Jesus doesn't play PWI. He may have suffered a lot for humanity, but he'd NEVER punish himself this way." -Abstractive | "I approve of bananas." -SashaGray0 -
Bump b:chuckle
1 Year has passed xP0 -
Elenacostel - Heavens Tear wrote: »Does Defense Level affect damage taken form DoTs? This would matter against a Phoenix, right?
DoT's Ignore defense level completely.
If a psy has 100 deff lvl and a nix bleeds you, you take direct damage as if u had 0 deff lvl ^^
That's why Psy's use Earth and Water DoT's against other Psy's or people with high Deff lvl, it damages them directly and is quite usefull when you nuke them and they try to heal.
A psy with 100 attk lvl won't deal more with DoT's I believe, the DoT's are based on base magical attack and on certain amount of damage over time~
b:bye
Hope that answers your questions, since no one else did ^^0 -
im getting 24 jades, +3 def cube neck, +5 def from r9 ring + 5 from lunar helm n robe, +3 def from r9 chest,+8 def from daily bless + 20def from r9 set bonus + seeker buf 3 def
thats 95 def lvl.should be 106% dmg reduction, means a tt boss that would hit me 10k will hit heal me when it hits?b:sad[SIGPIC][/SIGPIC]
I upload a vid of a rank 9 3rd cast seeker fight, 90 views.
I upload a vid of me in my boxers, 750 views.
I know what the PWI community wants.
b:laugh0 -
FlyRanger - Dreamweaver wrote: »im getting 24 jades, +3 def cube neck, +5 def from r9 ring + 5 from lunar helm n robe, +3 def from r9 chest,+8 def from daily bless + 20def from r9 set bonus + seeker buf 3 def
thats 95 def lvl.should be 106% dmg reduction, means a tt boss that would hit me 10k will hit heal me when it hits?b:sad
should hit you in the region of 1,5-2,5k'ish i guess. cos wite voodoo on full def lvl psy still get hit like 1k i think it was.. by beast in 3-3.I'm sorry for misspelling / mistyping and grammar b:surrender
102 - Archer - Heartz0 -
god damn, isnt it obvious that x def lvl doesnt = x% damage reduced? anyways...
im glad you guys are doin all the math behind it. really helped me a lot. but since you cant find the exact formula can someone do a test for me... (id do it myself but i dont have a friend with enough def lvls to try it out) find someone completely sharded with def +2 shards. k take off your weapon and all of there gear. hit them once. have them put a pice on (8 more def lvl) hit again. nother piece(8 def lvl) and so on and so forth. then put the jones blessing on the attacker and repeat. i really just want to see a graph with a line weather it be curved swurved looped dotted or circled....maybe ill ask UrDian...even though ive never talked to him before >.>0 -
RADD_RATT - Harshlands wrote: »god damn, isnt it obvious that x def lvl doesnt = x% damage reduced? anyways...
There is no more mystery and the correct formulas are at the top of the first post of this thread. These formulas are exact and have been confirmed over a very large range of values.[SIGPIC][/SIGPIC]
Refining Simulator - aster.ohmydays.net/pw/refiningsimulator.html (don't use IE)
Genie Calculator - aster.ohmydays.net/pw/geniecalculator.html - (don't use IE)
Socket Calculator - aster.ohmydays.net/pw/socketcalculator.html0 -
awesome...now if only i knew how to use those formulas. my last question would be if refining all youre gear with jade of steady defense the best way to go? or have jades and half vit stones? obviously it depends on class and all that. but im thinking the more resistances you have on your gear the more defense lvl will do for you, so where is the turning point?0
-
Socket Calculator: http://pwi-forum.perfectworld.com/showthread.php?t=9942020
-
damn...you archers know whats up! ill stop bothering you
and thaaaaank yoooou!!!!b:bye0 -
Yes, the archer forum is (in)famous for its maths.[SIGPIC][/SIGPIC]
ಥ_ಥ MOAR.
SkyKoC - How long is yours?0
Categories
- All Categories
- 181.9K PWI
- 697 Official Announcements
- 2 Rules of Conduct
- 264 Cabbage Patch Notes
- 61K General Discussion
- 1.5K Quality Corner
- 11.1K Suggestion Box
- 77.4K Archosaur City
- 3.5K Cash Shop Huddle
- 14.3K Server Symposium
- 18.1K Dungeons & Tactics
- 2K The Crafting Nook
- 4.9K Guild Banter
- 6.6K The Trading Post
- 28K Class Discussion
- 1.9K Arigora Colosseum
- 78 TW & Cross Server Battles
- 337 Nation Wars
- 8.2K Off-Topic Discussion
- 3.7K The Fanatics Forum
- 207 Screenshots and Videos
- 22.8K Support Desk