Zoner

You are completely ignoring the duration of corruption, it heavily dilutes the dps (and mana) loss from resists.

1000 corruptions with a 17% resist rate takes 1204 casts. 1000 of them have a cast time of 0. 204 of them have a cast time of 1.5 seconds due to waiting on GCD.

The duration of 1000 corruptions is 18000 seconds.

The initial cast costs 0 time with Improved Corruption.

204 resists adds 306 seconds due to recasts

Total time is 18306 seconds.

With a 0% miss rate the total dps time is 18000 seconds.

18306/18000 = 1.017

DPS loss from 17% resists for corruption is only 1.7%

On an added note taking improved corruption is an 11% dps increase for corruption since the cast time makes corruption take 20 seconds to do damage instead of 18.

 

deathmancer

1000 corruptions with a 17% resist rate takes 1205~ casts. Doesn't change the end result too much.

(x = 1000 / .83 to get the result, as you are trying to find x being the total casts to get 1000 to land)

edit:

This calculation doesn't really have any in-game application to the value of +hit.

Where will we ever have a situation where all we do is reapply corruption continuously each time it finishes and only that? In reality, if you've finished applying everything but corruption and corruption resists, you've now lost 1.5 seconds that you may have planned to use to shadowbolt or another dd ability, in which case the resist is more substantial in terms of potential damage. Almost every dot resist will ultimately result in either less drain-life time or shadowbolt or incinerate time for the most part on top of the damage lost for missing part of a tick of a dot in that time-frame.

 

MissMaldita

Well, I'm not sure the matter is as settled as you claim. From what I can gather on these boards, most people subscribe to the crits-can-miss theory.

It appears that this spreadsheet does too. It's calculating crits-can-miss from what I can tell.

 

Zoner

I agree but for all reasonable raid resist rates (sub 20%) the approximation is good enough. Misses in reality are functionally identical to mitigation, with cast time being multiplied by (1/(1-resistrate)). I work with graphics rendering so using approximations is a way of life for me if the results are close enough. I'll at least re-edit the posts and fix the numbers, since I'm stuck here at home today waiting on a delivery :P

 

TheOnly

(1 - 17%)= 83% of 1170 casts is 971.1 casts. If you cast corruption 1170 times with a 17% resist rate 1000 are not successful. There's a reason why the closed form of the infinite sum is the same as 1 / (1- resist rate). Its the same question as "83% of what is 1000". The answer is not 1170, that is the answer to "what is 17% more than 1000".



We are not concerned with the dps loss of corruption, its tiny. We are concerned with less shadowbolt casting time as a result. The spreadsheet is using these average cast times to figure out the remainder of the time leftover for the filler spells.

Not if you are chain casting. THats like saying that UA has a duration longer than 18 seconds. You can begin casting the second (non talented) corruption 2 seconds before the last tick goes off. Thus, when chain casting and refreshing DoTs, instant corruption is only a 0.5 second improvement.

THe DPS of the dots in this spreadsheet are calculated as the total damage divided by the adjusted duration. Adjusted duration includes a dot refresh unoptimal overlap constant you can adjust.
The "filler" dps is calculated by the remaining % of time left over not refreshing DoTs or life tapping to cast shadowbolt (or incinerate or drain life or drain soul or searing pain). The % to hit doesn't affect DoT dps that much because of the things you say (mostly). An extra 1.5 seconds every now and then on a 18 second duration DoT isn't a big deal. But it IS a fairly substantial effect for how many shadow bolts you can squeeze in between DoT refreshes, when you consider having 5 DoTs that can all miss.


Then, there is the effect mentioned just above where resists cause a "change in plans" that affects your refresh and may cause other DoTs to fall off for longer than expected. This is modeled with an average time that DoTs are not on the target due to such effects. You can experiment with this and add a second or three to the average "dots not on target" time and see how it affects dps. It hurts, but not nearly as much as decreased Filler spell cast time.

 

Zoner

That is a bit easier. 17% rate is about 1 in 6 instant casts needing a recast. So about 1 extra shadowbolt each every 180 seconds for corruption and unstable affliction. 1 extra shadowbolt every 150 seconds for immolate. 1 extra shadowbolt every 240 seconds for COA. 1 extra shadowbolt every 300 seconds for Siphon Life.

So in 300 seconds :
1.666 shadowbolts from corruption
1.666 shadowbolts from unstable affliction
2 shadowbolts from immolate
1.25 shadowbolts from COA
1 shadowbolt from siphon life

7.58333 extra shadowbolts could be cast in 300 seconds. But shadowbolt costs more mana than the dots we are replacing, so call it 7 bolts and 1 more lifetap to deal with the difference.

With a 1000 damage base theoretical dps is going to be around 850 for affliction. With that level of damage base shadowbolts are 1572, so 7 of those is at a non ruin 12% crit rate is 11660 damage added in 300 seconds, 38.86 dps. Assuming you believe 850 dps is right for full affliction in that gear, thats approximately a 4.57% dps increase overall going from 17 to 1% resist. Still a far cry from what +hit leads us to think will have us gain.

And for affliciton its all messy anyway because the dots get in and out of phase with each other, and you always end up with some dead time.

 

TheOnly

All of which, are testable in the spreadsheet.

From the spreadsheet:

900 shadow, 760 fire, 10% crit, 6% hit from gear (my current situation).

45/5/11 -- no imp CoA, no malediction, all other damage boosting affliction talents (4/5 suppression). Not counting life taps or pacts (so ~50 seconds dps).

DPS on a level 73 target.

Full DoT Cycle with CoA, no curses or shadow priests or fire mages = 841 dps on the spreadsheet.

This is assuming a 1.5 second DoT gap between refreshes on average due to non-perfect overlap, resists, and such.

+1 hit rating would give me 0.59 more dps.
+1 damage would give me 0.50 more dps.
1 more point suppresion gives me 7 dps. (0.83%)
2% more hit gives me 15 dps. (1.78%)
1 point devastation for 1% more crit gives 4 dps (0.48%)
CoD instead of CoA is 20 dps (2.37%)

% to hit increases dps a lot mostly because of fewer resists on shadow bolts and immolate, as the difference beetween suppression's 2% to hit on affliction and the 2% to hit from gear indicates.

Furthermore, an extra 2% crit is only 9 more dps. (1.07%)

Switching to 40/0/21 lowers dps to 825. (loss of 1.9%)

Switching to 41/0/20 increases it by 7 dps (+0.83%) (gain 5 crit from devastation, drop to 2/4 suppression and lose imp life tap -- if suppression is 1/5 or less, I lose dps but gain intensity and destructive reach)

What about the "dot gap" caused by not refreshing optimally?

If we are REALLY good and average 0.5 seconds per gap (essentially impossible), we gain 14 dps (1.66%).

If we are having a harder time with DoT refreshing and they average 2.5 seconds of 'gap' when they are not up, we lose 12.5 dps (1.48%).

If I am too focused on pew-pew shadowbolt action and the average time a DoT is not on the target climbs to 4 seconds, we drop to 811.6 dps, a loss of 29.5 ( 3.5% ).

To gain that 3.5% dps back from being lazy or inattentive, I need 67 more spell damage. Or an arcane elixir (cheapo) and adepts.


If my cast time lag gap is increased by 0.1 second, 26 dps is lost (3.1%)

Assuming a level 72 target, suppression and % to hit is already useless for me.
For a 73, hit rating is on par with +damage. About 60% of the % to hit worth is due to shadow bolts, and 40% of the worth is due to Dots.

Thus, if suppression gets you max hit for DoTs, hit rating becomes 60% as valuable as +damage.

Improved Shadow Bolt is responsible for 35 dps ( 4.16%).
Bane is responsible for 67 dps (7.96%)
If i went full devastation 5/5 for 5% more crit, it would be worth 23 dps (2.7 %), including the additional improved shadow bolt debuff uptime.

A few more notes:

* Shadow Bolts are roughly 60% of the cast time, and account for 39% of the damage ( 47%more if using CoS over CoA)

Using CoS over CoA loses 12.5 dps (-1.48%) but if I had Malediction, it would GAIN 7 dps (0.83%).

A CoS cycle gains slightly more from % to hit, and scales slightly better with + damage. With 0 suppression and 16% hit from gear, at 1500 shadow damage, a CoS cycle is more dps without malectiction or imp CoA, but imp CoA slightly overtakes it. CoD is of course better than both (slightly).

Back to my spec and stats...
Getting the improved CoA talent would gain 11 dps (+1.3%).
Since I play so often with a shadow priest around and thus use CoS, Malediction over Suppression is probably a good idea for me. Alternatively, full 5/5 shadow embrace isn't a bad idea since I know where to find those 4 points now (suppression).

All heroic bosses are level 72 anyway, only Kara + are there 73's to worry about.


So, in summary, Suppression and Devastation are BOTH pretty lame DPS boosters for affliction. % to hit gear is pretty good, primarily due to its effect on shadowbolt, and secondarily on its reduced DoT maintenance time. However it gets capped on non-bosses quickly.

% crit is pretty worthless compared to + damage or +hit per rating point. This is especially true because of the poor crit rating to crit chance ratio.

Ironically, the "profitability" of +damage and hit rating (value per point adjusted for item cost) is identical for affliction if %hit is not capped for both affliction and destruction spells.

If Affliction is capped due to suppression, but Destruction isn't the profitability is cut in half.

Shadow damage is 90% as good as +damage. It is slightly more profitable, item stat wise.

Lastly, delay between casts due to lag and timing have a much larger effect than "dot gap time". Spamming spells without delay is thus a much higher priority than getting the rotation 'just right'

0.1 seconds of spell cast delay causes as much dps loss as 2 seconds of 'dot gap'

 

Bolche

The spreadsheet asumes a 2 roll system for hit/crit/miss.
There is a lot of topics about this question on this forums (like this one).

a simplified exemple :
(without ruin)
SB average damage = base damage * (1 + 0.5 crit) * (1 - miss)

where a 1-roll system would lead to :
SB average damage = crit damage * crit prob + hit damage * hit prob

[top] base damage * 1.5 * crit + base damage * (1 - miss - crit)

base damage * (1 + 0.5 crit - miss)

Warlock DPS Spreadsheet

 

Neverlife

It would appear that the spreadsheet does not actually use the values of Siphon Soul (and Affliction Debuffs) at all.

While fiddling with the spreadsheet, I fixed that and I also added a section for HP/sec gained, which I find useful for fights with consistent AoE damage and for finding when I can be 100% self-sufficient.

 

Bolche

Indeed, I have added Siphon Soul and Affliction Debuffs but forget to add them in the drain life calculation. This will be fixed in the next version.

I don't know if I should include the HP/s in the main sheet, as I don't want to overload it with non-essential information. Well, I will add it in the next version as it can be usefull for aoe fights, as you said.

There won't be many new features in the 1.08 version. I now focus on fixing the bugs and improving the model rather than adding new features. Then I will probably add a new sheet tab dedicated to mana/life management (LT / DP / Potions / MP5 / OOM DPS / etc.). But if you have any suggestion to improve the spreadsheet, feel free to post it

Warlock DPS Spreadsheet

 

MissMaldita

I'd like to make a suggestion. Add a rank column for each talent and calculate the effect (for example 1.00) from that rank number. Using rank numbers would make it less easy to misidentify how many points have been specified for a talent. Right now, for example, some talents will display 0 for their effect when they have 5 talent points. This makes it easy to miss-configure a spreadsheet.

 

Demi9OD

The only other things missing are conditional values like crit % from demonic tactics and spell damage from Demonic Aegis and Demonic Knowledge. Easy enough to figure out on your own and just add to the add shad, add fire, and total crit % though I suppose.

Couple of questions, does anyone happen to have the base stam/int of a Felguard handy? I can figure out the others without respeccing. Also, what % of stam/int do pets receive from their master?

edit: Sorry was being lazy, found the scaling in another thread. Still would like to know base stam/int of Felguard.

 

• frmorrison

The spreadsheet assumes you will add in the spell damage/crit from Demonology talents yourself. I think that works fine, especially since each pet has different stam/int, so will change the spell damage gained (Felguard has the most stam/int on the pets).

The % of stam/int given to the pet is between 30-35%. If you mouseover stam/int on the base stats, it will say how much stam/int is going to the pet.

I guess if you wanted to know Felguard's base stats, you can work backwards (first remove 15% from each stat, then subtract the stam/int given to the pet by its master).

 

◊ Silmeria

Cool spreadsheet, I've been having a lot of fun using it, even discovering a few interesting things (maladiction'ed CoS being one of them).

However, I noticed the damage on incinerate does not change in relation to the usage of immolation or not. How is the damage bonus to incinerate during an immolation being factored in?

Sorry if I just missed something fundamental =)

 

Bolche

Yes, I agree some talents are really confusing, like Improved corruption (2, 1.6, 1.2, 0.8, 0.4, 0) where there 0 is the 5/5 talented one ^^
I'll probably do like in the Chalon/Pf rogue spreadsheet : 3 colomns for each talent : "talent name", "points" (from 0 to 5 for exemple), and the value used in the equation.


It asumes Immolate is alway up when casting Incinerate.

I could change it to use use the "use immolate" on/off switch (wich is not exactly correct since there could be an immolate from another warlock). But it is realy a bad idea to cast Incinerate without immolate.

Warlock DPS Spreadsheet

 

Bolche

1.08 out

* you can now enable/disable your filler spell (to test dot only dps)
* bug fix : contagion do not affect drain life and drain soul anymore
* added some comments on some unclear fields
* rewrite almost all talents to show a 0 to X list of choice (for exemple : 0,1,2 for nightfall)
* added HP/s information


I had to rewrite almost every equation due to the talent lists change. I checked everything twice, but please report any bug or error you notice.

I'm starting to gather informations about pets dps. The next version will contain a tab showing pet information (pet related talents, pet stats and dps)

Warlock DPS Spreadsheet

 

MissMaldita

Thank you for your hard work. Looking forward to seeing what you come up with for pets

 

Demi9OD

The numbers on incinerate seem much higher now, in the new version. Also, using searing pain as the filler spell has got to be broken, 3600dps seems a bit extreme :P

Upon a little more testing, Emberstorm seems to be the broken talent.

 

Bolche

thanks

Emberstorm fixed, you can download it again.

Warlock DPS Spreadsheet

 

Deathwing

Is conflag(or searing pain) working correctly? Both seem to vastly improve dps. I thought that in an immo/incinerate cycle, using conflag actually lowered your dps.

 

Crepe

According to some napkin calculations I did a while back, it should improve your DPS if your +Fire is less than 2000. This is due to incinerate's better scaling but conflag's higher base damage + casting time (1 GCD).

Keep in mind it tanks any mana efficiency you gained from casting immolate/incinerate vs. other destruction rotations.

 

Deathwing

Ah ha! I figured it out. The actual rank of S&F is multiplying the +damage for searing pain and conflag, making just anybody who would use those spells have wildly inflated amage values. Correcting the mistake gives me the proper result, that is conflag actually reduces your dps.

Otherwise, this is a very good spreadsheet, easy to use and even shed some light on new mechanics for me(full resists on dots not only costs damage but time as well).

 

cocidius