Arelenda

Thanks to everyone for clearing this up.

So if I got it right, Heroism/Icy Veins/haste are multiplied, boosting each other?
Cast speed = original speed / (1+ haste_rating/1577) / 1.3 / 1.2.

They obviously also boost any trinket or spellpower bonus you add. This is the major selling point.

Fire locks can use Flame Caps too, though. And mana gems and flame cap share cooldown if I'm not mistaking.

 

LockApologist

They do share a cooldown.

 

Rozzenwyn

Arelenda - To answer your question about making a new thread. Yes I think it would be greatly useful if we had 2 threads. One where you could update as new useful points are made and generally agreed to be "fact" and one for discussion. I for one have not read this entire thread because it is just too much. I read the beginning and have been trying to keep up for the past 15 or so pages with the new stuff that is being said but half of that is not really stuff I want to read until there is a clear consensus because I am not good at the number crunching and reading all the proofs just gets exhausting. Just my .02 cents since you asked.

Rozz

 

Torq

The reason Mages are a more valuable use of Heroism than Locks is due to the time compression effect of haste. Basically, with just heroism and IV, their 20 sec of +dmg from an on-use trinket becomes:

20 = t/1.3/1.2 => t ~= 31 sec

So roughly 30 seconds of on-use effect, over 20 seconds without their cooldowns stacked. Obviously, since they have more options available to use during that time period, they gain a much greater benefit than a lock, who would gain an effective 26 seconds of trinket uptime.

To be honest, looking at this, and back at the stacked+separate on the skull, I think for mages with an on-use+skull, their most effective gains would be to wait until after heroism+IV are off and pop the Skull right before their very next spell. I can't tell you for sure if that's the 100% min-maxed option, but that's just my estimate looking at it from here.

It boils down to the fact that trinkets and buffs are a static x seconds, while haste effects allow you to squeeze more spells into those x seconds in a very linear fashion. The more means you have to extend that x seconds to y seconds, the greater value those x seconds of uptime will have.

By the way, due to the linear nature of haste, you can model "effective uptime" on your trinkets as

Non-hasted time = Hasted Time/(haste effects)

Where non-hasted time would be the tooltip duration of the trinket, and "Hasted Time" would be equivalent to the trinket's uptime (amount of spells effected) with 0 haste. You technically never have to model haste on a per-cast basis; you can do it as a "sum of casts" using the total time to cast all spells as the "casting time." To go along with this, if you have x haste on your gear, and know the cast time of your spells with that amount of haste, you don't need the "base casting time" when calculating Heroism//IV, you can just use the hasted cast time and get the same results.

 

Shadowsnflame

I've recently specced into fire destruction for the first time, to test it out since the emberstorm changes in patch 2.4. I am trying to figure out what would be the best spell to cast at the tail end of an immolate. To the best of my knowledge, the most effective way to get the most out of your dots is to refresh them after they have gone their whole duration.

As immolate approaches it's last tick I usually have about 2 seconds left in which to cast, should I be casting another immolate and losing the final tick of the previous immolate? Should I be casting an incinerate which will not recieve the bonus from the immolate as it will be cast just after the immolates final tick, or should I be casting conflagrate and incuring the GCD before I refresh the immolate.

 

Torq

Depends entirely on how much haste you have... and consequentially, how far you are from your target.

With 0 haste, you can fit exactly 6 2/3 incinerates in an immolate; or 6 and 1.5 sec of excess. So, with 0 haste you can have a perfect rotation of immo -> 6x incin -> repeat; given that you have 0 travel time. See, travel time can throw this off by a large margin. Being max range from the target incurrs ~1-2 sec of travel time, meaning that last incinerate may actually land when the immo has ticked off, which, if I've read correctly, doesn't give that last incinerate the immo bonus.

To be entirely honest, I'd not even spec into conflag. It's not really worth a whole lot; you can always fill that small gap in your rotation with a lifetap if you're going to spend the GCD, and then just start your rotation over.

 

Arelenda

I used life tap for that. Never clip an Immolate. Especially in raids, just throw an extra Incinerate if LT is not an option. Chances are the other warlocks have Immolate up anyway.

 

LockApologist

Travel time is irrelevant. Spell damage is set the moment the spell leaves your hands. Thus, Immo falling off during transit doesn't matter.

See casting SB/Incin -> Casting CoS/CoE mid-flight. You do not get the damage bonus. Conversely, cast CoS/CoE -> SB/Incin -> Replace with different curse mid-flight. Damage bonus maintained.

 

Trickykid

It's not worth worrying too much about 100% immolate uptime. Worst case is you have one non-buffed incinerate for each immolate, which is 15/(15+incin_cast_time). Going from 100% uptime to that only drops my DPS by 0.6%. While that is still DPS, it's certainly not worth clipping immolate, or fussing too much with how many incinerates you can fit into 15 seconds. Just immolate when it's down, and incinerate otherwise. Haven't done the math, but lifetapping while immolate is down is probably a larger DPS loss since you don't have your higher DPCT spell up.

 

Torq

This isn't always the case. I know there are some examples of being able to land something while something else is in flight, and having the effect applied to both. I know I've had mages iceblock when a SB is in flight, and have them be immune instead of taking damage. This could be partly due to latency, but I don't think that's always the case.

It depends on how the immolate bonus is applied. If it's counted as a bonus to your spelldamage, I can see it being counted when the spell leaves your hand, but if it's considered extra damage from the target, it might be counted on land. I don't know, I'd have to do some logging or Dr. Boom testing to be sure.

It seems like on cast/on land stuff isn't terribly consistent across all mechanics, whatever the case is with immolate.

 

• frmorrison

WRT the iceblock, spell's getting +damage applied and actual damage taken are different. Spells and arrows must reach their target to do damage, however bonuses are applied at time of spell cast. So the ice block example is working correctly.

If you have an epic flying mount you can try to fly away and the spell will follow you, but only damage until you stop or slow down.

Immolate's bonus is just like other spell bonuses, it is counted at time of spell cast.

DK - Ashbane Failure is the condiment that gives success its flavor.

 

LockApologist

Having an effect apply to multiple spells (i.e. shatter combo or multi-spell reflect) and immunities are different than damage calculation.

And, in fact, the shatter combo supports what I said above, in that the target was frozen when the spells left the player's hands, so both get the crit bonus.

I agree that it's a bit odd that a player or mob can throw up an immunity shield mid-flight and have it take effect for that spell.

This is also very easy to test. Take off all gear, and throw immolate on a mob that something is tanking. VW will work. Cast an incinerate that will hit after immo drops. If the damage is over ~600, the affect applied. If it is under 500, it did not (barring partial resists). Above assumes you have 5/5 emberstorm. reduce dmg appropriately if not.

 

Alephnull

I have written a warlock gear optimiser in MATLAB that calculates the best gear set to achieve maximum DPS by wasting as few stat points as possible. It is based on a genetic algorithm which creates a 'population' of 10,000 gear and gem combinations (randomly) then selects the 8000 highest DPSing combinations, and 'breeds' them to create the next generation. It then adds 1000 random 'mutations' to the population by randomly changing gear options of random members of the population.

It will converge on the highest DPS output gear combination as defined by the file dps_func.m. This function finds the highest DPS, and considers everything _except_ drums of battle, heroism, totems of wrath, mana regen (its assumed lifetap scales with haste, plus if you are chain potting/dark runing, lifetapping is rare). All raid buffs, enchants and threads etc are added to the gear bonuses before the DPS calculation is carried out.

The gear database is in Script.m, it includes all decent DPS loot, as decided by me Its pretty much all cloth DPS from MH, BT, SW plus a few imba items like Quagmirren's Eye and the S4 arena dagger.

Chaotic skyfire diamond meta gem is assumed, because its amazing.

Proc items are calculated as the uptime fraction of the bonus stat. Eg. (15 seconds / 120 sec CD) * 320 spelldamage = 40 SD.

Files are available here as a RAR archive:
Free file hosting by Savefile.com

Run Script.m to begin the calculation. Note it will run until it hits 3000DPS, which it wont...So you will have to stop it after a few hundred iterations with Ctrl-C. It usually converges by 40-70 iterations. By limiting the variation of the population, it can be made to converge within 10 iterations, but this gives a less exhaustive check of the search space.

Any corrections are WELCOME. Please suggest items i have missed, and mistakes i have made. I will do my best to correct them.

Optimal gear setup currently appears to be :

Hood of the Malefic All sd meta
Amice of the Convoker All sd
Leggings of Calamity 2 sd 1 blue
Pendant of Sunfire All sd
Robes of Ghostly Hatred All sd
Belt of the Malefic All sd
Gloves of Tyris Power All sd
Bracers of the Malefic All sd
Boots of the Malefic All sd
Ring of Omnipotence
Loop of Forged Power
Grand Magisters Staff of Torrents 2 sd 1 blue
The Skull of Guldan
Quagmirrens Eye
Wand of the Demonsoul All sd
Nethervoid Cloak

To achieve 2785.2 DPS before drums, heroism, PI and ignoring lifetap time.

It also assumes you are an enchanter/JC...No tailoring or eng items have been included, because they will probably get trumped by Kil'jaeden loot. I could change this if people really want...

 

◊ Melbuframa

i would change it to include tailoring simply because the list that was posted a few pages back nets well over 2800 DPS in the spreadsheet, you included JC patterns yet not everyone is a JC

also Hex Head and M'uru trinket are a lot better then Q's Eye

 

Jakiri

Odd to see Quagmirran's Eye. How did you model the proc? You say you modelled proc items by bonus*duration/cooldown, but surely that only applies for on Use items?

 

Torq

I agree that it's weird to see that there. If you're modeling it with internal cooldown in mind, consider weighting the ICD as ICD + procchance*spellcasttime to get expected time to proc.

[E] If you remember my original modeling of the Blade of Wizardry, I got a higher value than it really had, because I was purely using duration/ICD, but I got a much more realistic one by using duration/expect TTP

 

Birgitte

*I haven't read the entire thread, so if it's been asked please over look me or point me to the right spot if you are familiar with the thread*
I have a question about the viable use of an Affliction warlocks in SW+ raiding. We are having a debate in our guild whether the utility of having one in the raids is worth the DPS loss to the individual player. This was started by an observation of one of our members that not many of the guilds killing Brutallus have Afflicition locks. Any information will be much appreciated.

Thank you.

 

Arelenda

Contrary to what you might have heard, Brutallus is pretty much a healer test. He dishes out incredible damage, and having Shadow Embrace is definitely worth it for progression. We had most issues keeping everyone alive. I'm not saying you don't need good dps, cause you definitely do. But it's not just about dps in that fight.

Optimal raid setup definitely includes an affliction lock at Brutallus.

 

Alephnull

The dps count that gets output will be lower than reality, due to the lack of drums and heroism. Im sure 2.8k would be easy with that kit...

Quagmirren's Eye is modelled as duration (6 sec) /( internal cd - duration ( 18 - 6) + time to proc) * bonus (320)

time to proc a 10% trinket is on average 10 casts, which would be 18-20 seconds when you have 400+ haste. So it works out as 6/(18-6+20)*320 = 60 haste in a constant nuke enviroment.

Basically, 18 secs is VERY short for an ICD, and 320 haste is loads....So my calculations suspect its actually pretty good.

Ill drop in the tailoring stuff and repost what i get for the output.

 

Iamminimal

I'm currently raiding in the higher raids (bt/hyjal, with za on the side n stuff)
I am 0/21/40 destro spec, sacrifice succy...

I have struck a point in the Alchemy Lab release on our realm (Barthilas, oceanic) which has put me in a spot of bother Q_Q

I think i have too much spell damage, not enough crit..and a little too less hit. if i were to get 202 hit (bout 170 atm) What stats are more important after that...ive started to get a few +10 haste gems..currently have 7, 3 gemmed and 4 in bank.

The World of Warcraft Armory is my current armory...If someone could please check what i need to upgrade the most..?

Tysm, greatly appreciated.

 

Nas

http://elitistjerks.com/f31/t10065-a...s_spreadsheet/

CE exalted ring, Crusade Card to replace the HMrT trinket, I'd go with a Scryer's Bloodgem over the Icon at your hit level, and I would go with Crimson Spinels over Quick Lionseyes once hit capped (in your belt for example). Some of the new badge reward pieces are no-brainers as well, the dagger and chest being a couple.

Keep in mind that I'm going with these suggestions blindly to be honest - they're just the obvious things. For solid numbers go through the spreadsheet I linked earlier and find out for yourself.

 

Krazen

This is ok with 7 healers. To be honest, though, do you think a guild that hasn't done Brutallus yet at this point has the dps to cover an 8 healer setup if one of your destro locks is gimping himself?

 

Torq

The last numbers I saw put it in the ballpark of a ~30 second ICD, same as the Blade of Wizardry. It's doesn't follow the "normal" rule of proclength*3.

And I'm not sure why you're doing (icd - duration) in the denominator; The icd starts when the proc activates (or that's how icds have been measured, anyway), so it should be 6/(30 + TtP) * 320

Where TtP would end up being your hasted speed without the proc, since that's the speed at which you'll be casting when the trinket is eligible to proc again.

 

Arelenda

Yes, I actually think a raid with Shadow Embrace will learn the fight quicker than one that doesn't. -5% damage is quite a lot, and the difference between affliction and destruction is a lot when we're talking personal dps, but compared to 28k raid dps it's not that impressive.

There is however, no way to quantify dps vs tank mitigation. It'll be subjective and raid dependent. You say tomato, I say infernal.


Update: 115 pages of this thread have been examined for useful links.Those links have been added to the compendium. I'll do the remaining pages soonish. If you think I missed any relevant posts, please let me know.

 

• frmorrison

Quagmirran's Eye internal cooldown is 45 seconds, so Hex Head is a better trinket.

6/(45-6+20)*320 = 32 haste is what you should enter.


It would be nice to see an LW/Tailor printout (add 20 haste for drums and the tailor chest).

DK - Ashbane Failure is the condiment that gives success its flavor.