nenad

*EDITED* Jun 2007, with new Warlock DPS XLS calculator

I tried to compare afflic, demo and destro builds. Main element that I tried to compare is damage on raid boss mobs between those 3 builds.

Originally I included only damage modifier talents in XLS calculations, but later I added other elements , so now its possible to set calculations for different:

1) Warlock base data
-Time period or fight duration
-items spell +dmg
-items shadow +dmg
-items fire +dmg
-crit% without talents
-added hit%
-spell haste %
-warlocks mana
-warlocks HP without talents

2) Enviroment data
-global cooldown delay+lag
-delay to recast DoTs
-raid buffs CoE / CoS
-raid buffs SW/Misery / Scorch
-additional shdw+fire buff mult.
-mob level compared to lock
-mob shdw/fire ressist
-Include Pet DPS
-sascrifice pet
-Pet type

3) Talent points (all DPS related talents in Affliction, Demonology and Destruction)


Goal was to require minimum changes in input if I want to compare DPS of same warlock with diferent talent builds, or same lock/build with different stats. Therefore, warlock data is entered without any talent modifiers (ie crit% without devastation/DT etc, so if i change from destro to afflic, i dont need to change warlock stats, only talent points). Also, frequency of all important spells are auto-calculated if that spell is chosen for use ... and that include needed LifeTaps, so when I change stats or build resulting in more/less mana usage, it will automatically adjust needed LTs and influence DPS.

In addition to DPS for chosen build, new XLS also calculate equivalents between +dmg stat and other stats like +shadow, +cri%, +hit and +haste.


New XLS can be found here:


Warlock DPS XLS calculator




Pictures and initial posts below are based on older versions of XLS calculations

 

nenad

some results from lvl 60 comparition - note that i made those before I added some elements like mob misses/ressists or HP/mana regens

guesstimate for warlock damage at lvl 60 over 100sec fight against average boss mob:



btw, one of main reasons for such advantage for affliction build is that I consider realistic scenario where lock need to regain mana in those 100sec (ie, not to use more than 6k mana, rest needs DP/LT) - and Dp > LT because it can be used in CD period.

But i made assumption that afflic lock can use DP 6 times in 100sec period. I guessed that you can DP 3-4 times before imp gets out of mana, and that imp need about 60sec to regen mana back.

comparative pies for afflic@60 related dmg % and cast time:



or



comparative pies for demolock@60 related dmg % and cast time:

 

Emily

Possible to delete my post here? (Didn't mean to break up a brother's flow!)

 

nenad

and finally some comparitions for lvl 70. So far without pictures ;p

For max DPS, with only limitation that you can not spend more mana that you have (ie, you must regain mana either with LT or DP), but without limitations on using your HP, ie you can LT as much as you need, and can expect some heal. This is viable only for short fights, or when you can count on healing:

81,656 dmg= 41/00/20 Full Afflic
75,865 dmg= 30/21/10 SM/DS
75,546 dmg= 31/00/30 DP/Ember
74,541 dmg= 07/41/13 Full Demo
72,576 dmg= 30/00/31 SM/Conflag
69,924 dmg= 00/21/40 SF/DS
68,698 dmg= 20/00/41 Full Destro

I tried even some non-41pts builds, and while most of them suprisingly did better than full Destro or full Demo, still full Affliction wins ;p

For no heals DPS, with limitation that you can not spend more than your total mana/HP, you can not count at outside heals, but still in 100sec period (ie you can end up with 100hp/mana at end of 100sec). This is viable scenario for most boss fights.

81,656 dmg= 41/00/20 Full Afflic
70,431 dmg= 07/41/13 Full Demo
68,452 dmg= 30/21/10 SM/DS
65,025 dmg= 20/00/41 Full destro

For sustained DPS, with limitation that you can not spend more mana/HP than you regen back, but without outside heals. For this calculations i presumed 1000sec period, and i will show DPS over period instead of total damage. This is mostly theorycraft, since you will rarely fight 10min in continuity without any heals and such:

785 DPS= 41/00/20 Full Afflic
591 DPS= 07/41/13 Full Demo
507 DPS= 30/21/10 SM/DS
495 DPS= 20/00/41 Full destro

While I can not be 100% sure in all formulas I used, above numbers definitelly take into account often ignored fact that you need to somehow get mana during fight and need to lose time doing it.

Since I'm personally demonology lock, I would appreciate if someone find and point out that i missed something in my XLS calculations , that would potentially close gap between demo and afflic ...otherwise I predict changing to afflic very soon ;)

 

◊ Praetorian

I'll let this stand separately because of the amount of content and the unique points it makes. Small questions or comments should be folded into larger threads, but this merits a thread of its own.

 

nenad

yes, I checked that thread (well, at least starting pages ;) , but assumptions there differ slightly from my assumptions:
- they do not take into account time needed for mana regen (DP/LT)- they have 18/19sec spell cycle period ... spell mix in that period does not represent raid boss fight, i wanted to fit all appropriate spells (+UA/siphon, LT/DP)
- zero +dmg gear was considered, and it has huge impact on lock DPS and differently affect DoTs, nukes and different builds

It does not means that calculations there are not correct - I just feel that for real raid boss fights my assumptions are more appropriate.

Of course, another reason why i didnt want to post this in other thread is that it would end up at page 62 - and not many ppl would find it there ;p

 

• frmorrison

Your numbers are not surprising, the dps of DoTs has been improved quite a bit, also assuming you have a Fire Mage @ 70 (15% more fire damage), Immolate/Incinerate could be decent dps.

Suggestion with regard to hosting, that host you used requires one to register, try savefile.com.

Know Thy group function; Thy DPS shall avoid aggro, Thy tank shall not attempt to DPS (Fixed in WotLK!), and Thy Healer shall not go AFK without notifying Thy group.

 

delljit

For what it's worth, at a little over 600dmg my imp gives 2 dark pacts before going oom.

http://www.rupture.com/characters/US...7Ganis/Fanatic

 

Emily

Can any lvl 70 locks give accounts of the dps of their pets? In fights where pets are given free reign to whomp the target, how much of the difference between Demo/Aff can that +30% damage Felguard make up?

 

nenad

tnx, I also got info from afflic locks that they can get 2 full and one partial DP before imp is oom. If that is 2.5 DPs per imp full mana, and imp regain mana in 60sec, that could still allow for 6 DPs in 100sec : 2 near start, and in remaining 90sec imp would regen 1.5 times mana bar, giving ~ 4 more DPs. But this is still guesstimate, so if some afflic lock have more precise answer to "how many DPs possible in 100sec", it would be nice ;p

BTW, I noticed slight error in XLS, related to improved Immolate at lvl 70 (when it was not present, i multiplied initial immolate damage with zero). Fixed this, included in new XLS that I hosted according to previous suggestion at another host:
Warlock damage.XLS

I will edit above posts to reflect new numbers - it increased numbers for afflic and demo only, since destro already had improved immol...making difference to destro even greater. I must admit that I expected destro to have at least same damage as demonology, or even greater. As it stands now, I either have another bug in calculations, or I had wrong expectations ;p

 

• Fres

Very interesting to read, but I have a few comments/questions-
[Edit: For your level 60 comparisons]Why did you exclude a traditional SM/Ruin build from your comparisons? In essence, does UA and the 5% bonus damage from contagion outweigh the dps increase from ruin and 5% to crit? What if we remove Curse of Agony?


Why did you assume any innate magical resists at all?
Why did you assume 0% to-hit from items?
Why the arbitrary 100 second duration?

 

nenad

btw, as comparition to above pie-charts, here is affliction @ lvl 70:

 

Shugorei

Why are you using Shadowbolt as the main attack for the detructolock, instead of an Immolate/Incinerate/Conflag cycle?

 

nenad

I have SM/Conflagrate build in there for lvl 70 comparition, which include SM/Ruin. As it can be seen, it does not outperform Full afflic.

In first chart (for lvl 60 comparition) you can see one graph for 'no CoA". Of course, it is easy to change it in XLS for lvl 70 builds and check, but end result is that , while we lower our DPS, gap between affliction build and rest is not lowered.

It was, as I mentioned in first post, guesstimate - I presumed that most boss mobs would have some ressist. Also, I wanted to be able to differenciate between builds with suppression and ones without, and to see how it affect differently nuke-heavy builds and DoT-heavy builds. As with all other stats, you can change it to 0 in XLS and check how it change results ;p

I was not sure that there is some item common to all warlocks that have +hit, but mostly since I didnt have any guesstimate I put 0 . What would be average value that we could expect warlocks to have?

As you said, it is arbitrary, but it is much closer to average boss fight than 18sec. Main reason was that in 100sec you can fairly accuratelly include whole number of DoT durations (XLS would accept partial durations in period, but i didnt want to use them). Allowing some 'slack' time, lag, waiting to finish current nuke, slow reflexes etc, I estimated about 6x Immolate and other 15sec DoTs, 5xCorruption and other 18sec DoTs, 4 x CoA at 24sec, 3x Siphon etc.

Main approach was to keep all DoTs up, and in pauses of recasting DoTs you use LT/DP as needed., and in all remaining time you cast nukes. But you can change spells that you want to use and how many of those in period in XLS and even use partial values, and you can change period also. I tried different spells distribution for every build and kept those that seemed to have best DPS results.

 

Brassicus

A lot of this number crunching theorycrafting has been done already, and not just for warlocks. Check it out for all casters here: http://shadowpriest.com/viewtopic.php?t=568

It suggests, under a couple different sets of conditions, that affliction and destro locks have very comparable damage, and that they both have equivalent damage to mages (assuming shadow priest present). Full demonology looks like ~10% less dps up to ~20% less as gear scales.

The author is really active, so if you poke holes in his assumptions, he'll update it. He might be a good one to talk to about additional theorycrafting for locks.

 

• frmorrison

Issues, if in a fight over 60 seconds, why not use Curse of Doom?

Like Shugorei said, why not use Imm/Incin/Conflag cycles rather than spam SB? Once you get that, assume 15% more damage (imp Scorch).

Issues with spreadsheet:
Wrong tooltip for Curse of Doom
Wrong tooltip for Drain Life
Wrong tooltip for Composites (You get 100 damage Fel Armor at 69, you get 50 damage at 62)
Bandaid is wrong amount, only 1360? I though it was 2400 for Heavy Netherweave. Also you can assume 20% more health gained with the bandaid from Fel Armor.

Know Thy group function; Thy DPS shall avoid aggro, Thy tank shall not attempt to DPS (Fixed in WotLK!), and Thy Healer shall not go AFK without notifying Thy group.

 

nenad

I calculate DPS for every build with 3 options for nukes: SB, Incinerate and SP .. then I use one with best DPS

And even for destro build SB resulted in better DPS, but that was very slightly better - almost same as using Incinerate. For other builds, Sb >> other two.

For example , for 20/0/41 destro build, i get in XLS:
a) Shadow Bolt 68,698
b) Searing Pain 65,155
b) Incinerate 67,334

while for 41/0/20 i get:
a) Shadow Bolt 81,656
b) Searing Pain 73,360
b) Incinerate 75,794

 

nenad

There is cycle, you can set it in XLS according to your wish, and in my calculations i used cycle of length 100sec.

For example, for afflic build i used in those 100sec: 19xSB, 4xLT, 6xDP, 4xCoA, 5xSoC, 6xImmo, 5xUA, 3xSiphon

I did not list exact order in which you will use those spells in those 100sec, but their cast times were accounted for. You can them any way you want within 100sec, as long as you attempt to recast DoTs as soon as they go out. For example:

CoA, UA, SoC, DP, Immo, DP, Siphon, SB,SB,SB .... or
UA, SoC, DP, CoA, Immo, SB, DP, SB, SB....

There is actually no special reason to state exactly in what order you use them for calculations, as long as you ensure that you cast above stated number of spells in that interval. And enough 'slack' was left to compensate for not being always able to immediatelly recast some DoT ( due to nuke being in progress etc). For example, UA&Corruption are 18sec DoTs, and I used only 5 in period, resulting in 90sec coverage ...you have 10sec slack time to recast either of those, where slack means you do damage by other spells like SB, not sitting around - although i included field in XLS for actual 'slack' ie doing nothing time even when you could be casting ( you can set that in 100sec due to lag, slow responses etc you waste 5sec for example)

 

spronk

Have you done a calculation for a more heavy affl warlock - say 46/5/0 at 60 - using drain life instead of sb? Modeling in the 4% sbolt proc from corruption & drain life of course. In that model drain pact is no longer needed, as you can lifetap in between drain life spam.

 

nenad

I tried using Drain Life for afflic build as replacement to SB, with full 60% bonus due to affliction debuffs, and it still result in significantly less damage than SBs, but I did not consider there additional Nightfall procs due to Drain Life, ie using SBs when it proc.

BTW, using LT is much worse for max DPS than using DP, since you can use DP during global cooldowns.

Anyway, even if i dont have exact numbers on chances of Nightfall proccing dur to both corruption and Drain Life, i optimistically estimated it at 25% in 5sec during which DrainLife is casted : 4% on damage, DL does damage every 1sec, Corruption 3sec, that is 6.66 'damages' during 5sec cast, resulting in abaout 25% chance. That is probably too optimistic since I do not remember such high ratio , but even with that....

I can squeeze 8 full DrainLifes in 100sec period and still have time to keep all previously mentioned DoTs ticking (that is bit under half time spent on draining, other time on DoTs and procced Sbs). If 25% of those drains would result in Nightfall, that is 2 times. Even if we ar further optimistic and make it 4 procs, that result in 4 'insta' cast Sbs, but those ar enot exactly insta considering 1.5sec cooldown.

With all that in consideration, ie with 8x Drain Life, 4x Instacasted SB and DoTs as mentioned before, i calculate 73,412 damage, which is less than damage using only SBs (81,656)

About only serious advantage of DrainLife in that scenario is that I gain about 13k HP during period, as opposed to losing 373 HP if only using Afflic ;p

Basically, for max DPS i presumed that it is best to use spells that have highest damage per cast time ratio. For example of 41/0/20 afflic, that looks like:

 

• Nfariessence

I've seen you say this a few times in this thread. Does Dark Pact not invoke the Global Cooldown anymore? I thought I was getting "That ability is not ready yet" in Nagrand this morning...

 

nenad

because CoD has lower DPS than CoA (even more for destro lock than for afflic). In "sustained DPS" i used CoD since it has much better damage/mana.

For destro build, i used both Immolate and Conflag, and then compared DPS if SB or Incinerate are used as fillers. What additional damage i would have in case of using Incinerate , ie where those 15% would came from? Maybe I missed something in that part ;p

What values are wrong? I checked most data from thottbot
BTW, both of those spells are not used in max DPS cycles, so it wont change numbers, but would be nice to fix if i missed something.

fixed that ;p But of course, it did not change numbers, since it counted 100 for lvl 70 already

fixed that, but again bandaid was not used in any cycle;p

 

nenad

Well, I must admit it was some time since i had afflic build, and i think that i could use DP even while casting other spells. That indicated DP being on pets CD. But based on info that i got now, i changed DP back to have global cooldown. I only wonder if anyone remembers time when it was on pets cd ;p

 

Pestilence

I am curious about few things

Level 60 comparison

Why pick 0/41/10 for raid DPS build? and not 9/21/21. Did you include FG DPS in your calculation?
Same thing for Destro? no soul leech / backlash
Why stack crit if you 41 affliction? you benefin alot more from pure +damage then crits with DoT heavy build
Did you also look at 30/0/21 and take imp SB procs? 20% crit or so buffed?
Was CoS already on the mob?
It would be nice to see what builds you used while gathering the data.