Please upgrade your browser for the best possible experience.

Chrome Firefox Internet Explorer
×

Underworld PvE Relics

STAR WARS: The Old Republic > English > Classes
Underworld PvE Relics

dipstik's Avatar


dipstik
04.18.2013 , 02:19 AM | #1
underworld stuff: rememebr to check if the relic is force or tech! vendor has no tech internal/elemental relics, so tech users with less than 38% armor reduction (see notes at end) shuld wait for bioware to address this and get the next BiS relic.

assumptions need to be made. lets say dot specs have 1 action every 0.7 seconds and everyone else has 1 action every 1.5 seconds. lets say if you have a 30% chance to proc something, that it will take you 3 actions to proc it after the cooldown of the proc.

on use power: 425*1.05*0.23= 102.6 bonus damage per action. for 30 seconds: dot users get 42.9 actions, for 4403 damage every 120 seconds giving 36.7 dps. other users get 20 actions for 2052 damage over 120 seconds gives 17.1 dps.

proc power: 550*1.05*.23= 132.8 bonus damage per action. for 6 seconds: dot specs get 8.57 actions, for 1138.3 damage every 20+3*(0.7)= 22.1 seconds giving 51.5 dps. other users get 4 actions for 531.2 damage every 20+1.5*3= 24.5 seconds giving 21.7 dps.

proc elemental/internal: 264 damage each pro. dot users: one proc every 4.5+3*0.7=6.6 seconds gives 40 dps. other users: proc every 4.5+3*1.5=9 seconds for 29.3 dps.

proc kinetic/energy: assuming the equations below hold and data from guildy was good (275 damage form 352 kinetic attack), and my assumption that arsenal stance still has 35% armor reduction from stance and 20% from tracer stack, the the following holds: operation dummy has 6354 armor rating, which means you need 38.1% armor reduction to make the kinetic relic do as much damage as the internal/elemental relic.



armor reduction: ar(p):=x*(1-p) / ( x*(1-p) + 200 * 55 + 800 ), x is armor rating, p is armor reduction (remember the amount from skills vs. debuffs add multiplicativly... so your total will be p=1 - (1-p1)*(1-p2).. where p1 is from debuffs and p2 is from skill/stance etc.
i*(1-ar(p)) = f, where i is intial kinetic/energy damage, f is final damage after armor reduction

Ailaria's Avatar


Ailaria
04.18.2013 , 03:31 AM | #2
I'm afraid you've completely forgotten about both damage coefficients and any spec-related bonuses, which makes your numbers for the power relics completely wrong, especially the part about dot specs. The damage proc relic numbers are also probably wrong, since they benefit from your stats to some extent as well (after all, why else would force-users prefer the force relic etc.).

In other words, a nice effort, but the calculations are more complicated than what you've done there.

dipstik's Avatar


dipstik
04.18.2013 , 09:57 AM | #3
Quote: Originally Posted by Ailaria View Post
I'm afraid you've completely forgotten about both damage coefficients and any spec-related bonuses, which makes your numbers for the power relics completely wrong, especially the part about dot specs. The damage proc relic numbers are also probably wrong, since they benefit from your stats to some extent as well (after all, why else would force-users prefer the force relic etc.).

In other words, a nice effort, but the calculations are more complicated than what you've done there.
not sure what you mean by damage coefficents. like (1+c*s)*(1-miss)? if this is the case, then the above analysis is still fine because you are comparing bonus damage to bonus damage. all of the dps numbers (just used to compare not predict parses, as a simulator is the best thing for that, and i agree the simulator is much mroe complicated) will have (1+c*s)*(1-miss) apply. even the force/tech internal/elemental has resist now.
which spec related bonuses would you like to add to the model?

i only ask because i think you are trolling me based on your post count

mastirkal's Avatar


mastirkal
04.18.2013 , 10:57 AM | #4
He's not trolling you, the coefficent of DPS is as followed if we use lightning strike as an example

SpellDamage: SpellType=>Force, StandardHealthPercentMin=>0.14, Slot=>None, Coefficient=>1.5, StandardHealthPercentMax=>0.16, DamageType=>Energy

Expected damage=((base damage+(Coefficient*Bonus damage))*Expected modifier) (Then any buffs and debuffs including the % to even hit.)

how you calculate this is your low and high modifier (o.14-0.16 x 7055/2 is the base Damage amount. We then apply your bonus damage x the coefficient to add to the damage of the base amount. Once that is done we apply the overall expected modifier which is your surge and critical rating expected on top of your power based DPS, but can be chucked up to =(1-(crit/100)) +( (1+(surge/100))*(crit/100)) on top of that amount to get your expected DPS for that one move before adding any buffs like bonus damage or higher crit rating for certain moves, but not all.

Aka, you need more data then just doing simple math to figure out how much these proc relics work. Non DoT specs BiS is the new Serp Assult and power clickie.

dipstik's Avatar


dipstik
04.18.2013 , 11:38 AM | #5
Quote: Originally Posted by mastirkal View Post
He's not trolling you, the coefficent of DPS is as followed if we use lightning strike as an example

SpellDamage: SpellType=>Force, StandardHealthPercentMin=>0.14, Slot=>None, Coefficient=>1.5, StandardHealthPercentMax=>0.16, DamageType=>Energy

Expected damage=((base damage+(Coefficient*Bonus damage))*Expected modifier) (Then any buffs and debuffs including the % to even hit.)

how you calculate this is your low and high modifier (o.14-0.16 x 7055/2 is the base Damage amount. We then apply your bonus damage x the coefficient to add to the damage of the base amount. Once that is done we apply the overall expected modifier which is your surge and critical rating expected on top of your power based DPS, but can be chucked up to =(1-(crit/100)) +( (1+(surge/100))*(crit/100)) on top of that amount to get your expected DPS for that one move before adding any buffs like bonus damage or higher crit rating for certain moves, but not all.

Aka, you need more data then just doing simple math to figure out how much these proc relics work. Non DoT specs BiS is the new Serp Assult and power clickie.
ah okay, so each ability has its own coefficent for bonus damage, so

dps=(base+bd*coeff)*(1+c*s)*(1-miss)/(1-a)

where base is base dps, before bonus damage, crit, surge, alacrity or accuracy, but after skill tree stuff that doesnt affect those latter values.
bd is bonus damage
coeff is the multiplier of bonus damage for each ability
c is crit, fraction
s is surge, fraction
miss is chance to miss, fraction, based on accuracy and offhand info
a is alacrity, fraction


it seems difficult to seperate out the coefficent data. the only thing i can think of is analyzing 2 parses with different amountds of bd and analyzing the non crit values.

low=base+bdlow*coeff
high=base+bdhigh*coeff

since bdlow and high are known, as well as low and high, we can solve for coeff and base.

to get a friendly term to use before crit, surge etc, get each abilities contribution to dps, as a fraction, call it f1 through fn. solve for the base values (final=(base+coeff*bd)*(1+c*s)) . use a weighted average:

base=f1*base1+f2*base2...+fn*basen. because f1 is based off number sthat include surge etc, this is an approximation. we can do something similar for the coeff,

coeff=f1*coeff1+f2*coeff2+...+fn*coeffn

this should hopefully give us

dps=(base+coeff*bd)*(1+c*s)*(1-miss)/(1-a)

where we can use weighted sums for crit, surge, alacrity and miss. this is just to avoid rotation simulations.

dipstik's Avatar


dipstik
04.18.2013 , 08:16 PM | #6
here is how you find the values. might be easier if you remove all yoru skill points to remove as many buffs to the abiltiy as possible

hit1a=min*base+bd1*coeff
hit1b=max*base+bd1*coeff

hit2a=min*base+bd2*coeff
hit2b=max*base+bd2*coeff

where bd1 and bd2 are bonus damage levels, hit1a hit 1b hit 2a and hi2b are non crit damage values, one high and one low for each parse.

min max base and coeff vare the values we are looking for

some equations work out nicely

coeff=(hit2a+hit1a)/(bd1+bd2)

min = (hit2a*bd1-bd2*hit1a)/(hit2a+hit1a)

base = -(hit2a+hit1a)/(bd1+bd2)

max= ((-hit1b+hit2a+hit1a)*bd1-hit1b*bd2)/(hit2a+hit1a)

. for a system with no min max you can use coeff and base equaitons, just remove a and b suffix.

Ailaria's Avatar


Ailaria
04.19.2013 , 06:49 AM | #7
Quote: Originally Posted by dipstik View Post
not sure what you mean by damage coefficents. like (1+c*s)*(1-miss)? if this is the case, then the above analysis is still fine because you are comparing bonus damage to bonus damage. all of the dps numbers (just used to compare not predict parses, as a simulator is the best thing for that, and i agree the simulator is much mroe complicated) will have (1+c*s)*(1-miss) apply. even the force/tech internal/elemental has resist now.
which spec related bonuses would you like to add to the model?

i only ask because i think you are trolling me based on your post count
As was pointed out, I'm not trolling, the post count is due to my general laziness in even reading the official forums.

Anyway, the points I was trying to convey were that a) you cannot compare relics with power to the damage proc relic without full theorycrafting on the given spec, since you need to know exactly which abilities are used and how often due to the different bonus damage coefficients, and b) the comparison of dot-users vs. others is wrong for similar reasons, since dots naturally have very low bonus damage coefficients for each tick.

In fact, the only comparison like this that you *can* make without having a lot more data is the comparison of proc power vs. on use power, and there your approach does seem correct.

Apologies if I was unclear before.

dipstik's Avatar


dipstik
04.19.2013 , 11:42 AM | #8
Quote: Originally Posted by Ailaria View Post
As was pointed out, I'm not trolling, the post count is due to my general laziness in even reading the official forums.

Anyway, the points I was trying to convey were that a) you cannot compare relics with power to the damage proc relic without full theorycrafting on the given spec, since you need to know exactly which abilities are used and how often due to the different bonus damage coefficients, and b) the comparison of dot-users vs. others is wrong for similar reasons, since dots naturally have very low bonus damage coefficients for each tick.

In fact, the only comparison like this that you *can* make without having a lot more data is the comparison of proc power vs. on use power, and there your approach does seem correct.

Apologies if I was unclear before.
i apologize for being so dismissive, as i am the one being ignorant. i basiclly used an analysis someone did when DG came out, assuming they knew what they were talking about, but they may have been comparing static power to on use power.

without a simulator, doing a comparison is going to be a long trudge through unfamiliar waters. best bet is to just parse with one, then the other, using the 2 hour turn in timer. problem is the difference in dps could be 5, which will not be noticeable with a small sample of parses.

Yngow's Avatar


Yngow
04.22.2013 , 10:33 AM | #9
So boundless + proc power looks like bis for dps? I mean, that was my assumption from the beginning when I saw the new relics but I was hoping for something else, I like diversity and having a best choice for each spec/class instead of just a no-brainer for everyone, there's just no reason for the proc dmg relics to be in game if that is case.

KeyboardNinja's Avatar


KeyboardNinja
04.22.2013 , 03:54 PM | #10
Quote: Originally Posted by dipstik View Post
underworld stuff: rememebr to check if the relic is force or tech! vendor has no tech internal/elemental relics, so tech users with less than 38% armor reduction (see notes at end) shuld wait for bioware to address this and get the next BiS relic.
Why? Tech Elemental *does* use the Tech crit chance, while Force Internal uses the Force crit chance (verified via parsing).

Based on your analysis, as well as the on-use vs proc power analysis, I would say the relic choices are pretty much what they used to be: internal/elemental proc + on-use power unless you are a gunnery commando/arsenal merc, in which case you want the kinetic/energy proc + on-use power. If you absolutely cannot bring yourself to time an on-use power relic, or you have a spec without any significant burst phases, it's probably better to use the on-proc power relic.
Computer Programmer. Theory Crafter. Dilettante on The Ebon Hawk.
Tam (shadow tank) Tov-ren (commando healer) Aveo (retired sentinel) Nimri (ruffian scoundrel)
Averith (marksman sniper) Alish (lightning sorcerer) Aresham (vengeance jugg) Effek (pyro pt)

December 13, 2011 to January 30, 2017