Most effective warlock raid dps spec? - Page 7

Zellyn · 07/31/06, 2:27 AM

Originally Posted by Hangman

- Zibro ,
I lose the screenshots from ct_combatstats from that fight :(

Combat Stats isn't going to give you an accurate picture of your DPS. It measures DPS over a 30 second period. It won't give you an accurate idea of DPS. Stats on individual actions, sure, but not DPS.

Unless CT_Combatstats is different from the CombatStats they've been bundling into CTMod for the last six months, I guess.

Kobal · 07/31/06, 5:07 AM

Something else I have wondered, and what could be a nice theorycrafting exercise: What would be the optimal Warlock Raid-DPS build under the following boundary conditions:

- You must have an Imp out at all times
- You must have the 3/3 Improved Imp talent and the 2/2 Improved Healthstone talent.
- The fights are not particualar aggro sensitive, but you also can not go completely wild on aggro (no Broodlord, but also no Patchwerk).

The only candidate I could think of was a heavy destruction build like 7/10/34. Have I missed something?

Christmas · 07/31/06, 7:05 AM

Originally Posted by Azamad

The MD/DS Trick to get both bonuses is indeed considered a bug. It is included in the known issues list on the official forums. Just scroll down a bit until you reach the warlock section.

http://forums.worldofwarcraft.com/th...p=1#post148782

That post also has this bug:

# Pets
# The Succubus can despawn before it can be resurrected.

Which is rather counter-intuitive.

Mosh · 07/31/06, 8:20 AM

Originally Posted by Moleva

I don't see how it's any different.

The longer the uptime the greater benefit it provides to all warlocks.

For my simple 2 warlock system, say that one warlock has a 10% crit rate and the other has a 15% crit rate. The uptime for the debuff is 65.8% of the time. If there was only the 10 crit warlock, the uptime would be 34.39%, and for only the 15 crit warlock, the uptime 47.8%. Now obviously the 10 crit warlock gets a greater benefit from having the 15 crit warlock there than vice versa, but the 15 crit warlock gets a fair increase to his dps none the less. And this is at a pretty substantial difference in gear. With both at 15 crit, the uptime is 72.75%. That's a huge gain for both warlocks.

This is wrong.

Uptime is 1-(1-avgcrit)^4, avgcrit being the average critrate of all Warlocks in the raid. In the 10/15% case assuming both Warlocks have an equal manapool and the same spec with regards to mana usage, the uptime will be 41.4%

• bartolimu · 07/31/06, 10:31 AM

Originally Posted by Kobal

Something else I have wondered, and what could be a nice theorycrafting exercise: What would be the optimal Warlock Raid-DPS build under the following boundary conditions:

- You must have an Imp out at all times
- You must have the 3/3 Improved Imp talent and the 2/2 Improved Healthstone talent.
- The fights are not particualar aggro sensitive, but you also can not go completely wild on aggro (no Broodlord, but also no Patchwerk).

The only candidate I could think of was a heavy destruction build like 7/10/34. Have I missed something?

In that case I'd probably say something on the order of 30/5/16 - SM/impHS/Shadowburn, with 5 points in Devastation. Frankly I'd rather see the 2 points go into Destructive Reach or even Intensity rather than Improved Healthstone.

Heavy destruction isn't good in aggro-limited fights, and it's not generally as efficient as more Shadow-centered builds. You can never count on Immolate hanging around for even close to the full duration, which it needs to if Conflag is going to be of much use.

In other news, I respecced to SM/Ruin yesterday before running Abomination wing. My place last week on DMs? 19th. My place this week on DMs, despite a respec, using Demonic Runes and Major Rejuves to cut down on cooldowns, and hitting Shadowburn every 15 seconds or so? 19th. No event parse because I forgot, but I promise you it's not because I suck at playing. Even using every optimal consumable, there's just too much damage inherent in warlock damage.

I guess I'll try major mana pots next time and see if they change anything.

Kytrarewn · 07/31/06, 11:07 AM

Originally Posted by bartolimu

]
Frankly I'd rather see the 2 points go into Destructive Reach or even Intensity rather than Improved Healthstone.

That's something that's been confusing me, why everyone feels the need to go for 2/2 iHealthstone. I've always liked having numerous locks with different flavors of healthstones during a given raid.

Moleva · 07/31/06, 12:17 PM

Originally Posted by Mosh

This is wrong.

Uptime is 1-(1-avgcrit)^4, avgcrit being the average critrate of all Warlocks in the raid. In the 10/15% case assuming both Warlocks have an equal manapool and the same spec with regards to mana usage, the uptime will be 41.4%

Ah, you're right. I forgot about charge consumption. That kinda stinks then.

NiXXeD · 07/31/06, 12:52 PM

Originally Posted by Moleva

Originally Posted by Mosh

This is wrong.

Uptime is 1-(1-avgcrit)^4, avgcrit being the average critrate of all Warlocks in the raid. In the 10/15% case assuming both Warlocks have an equal manapool and the same spec with regards to mana usage, the uptime will be 41.4%

Ah, you're right. I forgot about charge consumption. That kinda stinks then.

If the numbers are true, and say you're doing 500 DPS normally, that means that 41.4% of the time you're doing 600 dps passively. I'd say that's an improvement...

Over 1 minute of doing 500 dps without Imp SB you'd have 30k damage
Doing the same WITH imp SB would yield 32.5k damage which comes out to be about an 8.33% improvement in DPS. I'd say that's noticeable for a tier1 talent. But this all increases and scales with gear... The more crit your locks stack, the higher % the proc stays up, leading to a higher overall increase in the warlock damage.

With all of what's said, I don't think I'd ever give up ruin, and the typical Destro tree talents... They make noticeable differences in every build, and it's just too required to my playstyle. I've tried dropping them at times, but never successfully, as I miss them too much.

Oth · 07/31/06, 2:22 PM

Originally Posted by Kytrarewn

Originally Posted by bartolimu

]
Frankly I'd rather see the 2 points go into Destructive Reach or even Intensity rather than Improved Healthstone.

That's something that's been confusing me, why everyone feels the need to go for 2/2 iHealthstone. I've always liked having numerous locks with different flavors of healthstones during a given raid.

The goal is to minimize the wasted demonology points for an MD or DS build.

For my 7/21/23, for example, I only have one 'junk' point (in intellect, I think). Any point moved away from Imp. HS is a point going into something useless (the stat talents are far inferior to the pet buff talents for both PvE and PvP).

Multiple-rank healthstones are better implemented if you go to MD/SL, because you have four points to sink and at that point moving one away from Imp. HS isn't so bad.

Auphi · 07/31/06, 2:46 PM

Patchwerk DPS Spreadsheet
http://pages.nyu.edu/~xc228/xDPS.xls
It currently only takes into account Shadowbolt DPS. I don't mean for the Spreadsheet to be used as an predictor of how much DPS you will do, but rather as a way to rank different approaches to a Patchwerk-like scenario. (Where healing and long-term DPS are at a premium)

The spreadsheet is part of a larger workbook, so the "Stats" section corresponds to my Warlock. If you want to adjust it to your gear, use the "Test Stats" section to modify the values. I did allow for the option of Imp Shadowbolts, following the law of mediocrity, if you feel I overvalued or undervalued it, feel free to adjust the modifiers accordingly.

Skav · 08/01/06, 6:42 PM

Hello there (first post here, pardon the mistakes as I am French)

I wrote a simulator some times ago to study this kind of questions :
http://www.skav.org/theorycraft/warlock_simulator.php
It doesn't include all trinkets or regeneration strategies, but should give a good approximation.

My results for "patchwerk-like" (but over 10 minutes) DPS : http://www.skav.org/theorycraft/warlock_study_regen.php

(I also did some simulations on improved shadowbolt : http://www.skav.org/theorycraft/warlock_impsb.php )

Christmas · 08/02/06, 4:52 AM

Originally Posted by Skav

Hello there (first post here, pardon the mistakes as I am French)

I wrote a simulator some times ago to study this kind of questions

Fantastic simulators. I'm having great fun playing around with upgrades on them. Thank you! I have a newbie question though: the mana regeneration rate (given as 5 per tick, default) - is that some rate that is true for all warlocks within the 5SR? Or do I need to set it to zero?

Or, is that a base regeneration rate, which is effectively ignored by the simulator because all the time is spent in the 5SR?

balorina · 08/02/06, 6:03 AM

Originally Posted by Skav

Hello there (first post here, pardon the mistakes as I am French)

I wrote a simulator some times ago to study this kind of questions :
http://www.skav.org/theorycraft/warlock_simulator.php
It doesn't include all trinkets or regeneration strategies, but should give a good approximation.

My results for "patchwerk-like" (but over 10 minutes) DPS : http://www.skav.org/theorycraft/warlock_study_regen.php

(I also did some simulations on improved shadowbolt : http://www.skav.org/theorycraft/warlock_impsb.php )

Your simulator is coming up with numbers that are too low. It is in how you calculate average shadow bolt damage and filtering down from there.

Vytae · 08/02/06, 7:06 AM

Id have to go with Sm/DS (been it for almost a year). After reading this thread i respecced to sm/ruin (like the good old days!) for a night. I played a few BGs,and forgot how much i missed faster shadowbolts,destructive reach and shadowburn. I also managed to pull aggro on every second boss,and basicly every trash mob i targeted. Luckily i warned my raid that i had repspecced and couldnt be held responsible for ANYthing ;) I could manage my aggro fairly enough,but those crit spikes kept killing me not to mention without VW or Fel sac my dps went out the window after half a minute if aggro was a non-issue. Trash survivability sucked without that extra 1khp too.

Sm/Ruin still reigns supreme as best over dps spec, but its the epitomy of everything thats wrong with the warlock class. Horrible mana efficiency,Horrible Aggro managment and half-assed surviability.

Ive come to realize that Demonic sacrafice is almost required these days for warlocks raiding simply because it allows us to customize ourselves to any given fight to shore up our glaring weak spots.

Skav · 08/02/06, 11:43 AM

The mana regeneration is what you can get from buffs, mp/5s on items or enchants, I consider (which is mostly true) that a warlock spends most of the time inside the 5 second rule interval.

Originally Posted by balorina

Your simulator is coming up with numbers that are too low. It is in how you calculate average shadow bolt damage and filtering down from there.

Care to explain ?

Btw, I assumed that hit and crit are independant rolls, ie crits can be resisted and don't work like melee. I couldn't find any reliable information about the question.

Aghama · 08/02/06, 12:33 PM

Originally Posted by bartolimu

As Greybone already points out, the loss of Bane is extremely difficult to counter. I ran some numbers a few months ago and found that SM/DS never scaled up to be better than SM/Ruin - not even at +800 damage. If I can find that post again I'll quote it here, but it was on another board in a thread that's likely archived.

I can't believe I read this 6 months ago.

SM/Ruin Shadow Bolt Casting Time: 2.5 sec
SM/DS Shadow Bolt Casting Time: 3.0 sec
Assume 15 seconds of casting.
SM/Ruin: 6 bolts
SM/DS: 5 bolts
That's a full "cycle" in that it shows the full advantage of Bane but doesn't make the numbers any bigger than they need to be.

SM/Ruin Shadow Bolt Damage Formula:

    ([(base damage) + (.85 * [+damage total])] * 6 * [SM bonus]) * (1 + [crit%])

SM/DS Shadow Bolt Damage Formula:

    ([(base damage) + (.85 * [+damage total])] * 5 * [SM bonus]*[DS bonus]) * (1 + [crit%/2])

Obviously crit is much more helpful to Ruin, which we already knew. Question is, at what point does crit become better long-haul DPS than +damage? Or, to put it differently, what %crit is necessary for SM/Ruin to be stronger damage overall than SM/DS?

Assume 100 +damage.

    ([(481) + (.85 * 100)] * 6 * 1.10)  * (1 + x) = ([(481) + (.85 * 100)] * 5 * [1.10] * [1.15]) * (1 + [x/2])
    3736 * (1 + x) = 3113 * 1.15 * (1 + [x/2])
    3736 * (1 + x) = 3580 + (1 + [x/2])
    3736 + 3736x = 3580 + 1790x
    3736 + 1946x = 3580
    1946x = -156
    x = -.08

So, contrary to my initial thoughts, SM/Ruin beats out SM/DS at +100 damage. An SM/DS warlock would need about 8% more +crit than an SM/Ruin lock to deal more sustained DPS. Is that true at higher +damage levels, though?

Assume 400 +damage, which is about midway between my everyday set and my DPS set in 1.9 (and what I tested with):

    ([481 + (.85 * 400)] * 1.10) * 6 * (1 + x) = ([481 + (.85 * 400)] * 5 * [1.10] * [1.15]) * (1 + [x/2])
    5419 * (1 + x) = 4515.5 * 1.15 * (1 + [x/2])
    5419 * (1 + x) = 5193 * (1 + [x/2])
    5419 * 5419x = 5193 * 2597x
    5419 * 2822x = 5193
    2822x = -226
    x = -.08

Motherfucker. I just hate it when I'm wrong. In fact, the effect doesn't seem to scale at all. SM/Ruin is better at all levels of +damage without a substantial increase in spell crit (which is even harder to achieve thanks to Devastation).

Note, however, that this ONLY takes into account Shadow Bolt spam. The 15% extra damage to Corruption, Siphon Life and (if you can use it) CoA would probably make up for a substantial portion of this shortfall. I don't have time to work those into the equations right now, but some day I will.

Math nerds, check my math for algebra or simple computational errors. I don't think there are any but hey, I've made stupid mistakes before.

Emarius · 08/02/06, 1:04 PM

Originally Posted by Christmas

Originally Posted by Azamad

The MD/DS Trick to get both bonuses is indeed considered a bug. It is included in the known issues list on the official forums. Just scroll down a bit until you reach the warlock section.

http://forums.worldofwarcraft.com/th...p=1#post148782

That post also has this bug:

# Pets
# The Succubus can despawn before it can be resurrected.

Which is rather counter-intuitive.

Not necessarily counter-intuitive. Its possible they intend for pets to be resurrectable but for the DS buff to fade when the pet becomes active again.

balorina · 08/02/06, 1:13 PM

Btw, I assumed that hit and crit are independant rolls, ie crits can be resisted and don't work like melee. I couldn't find any reliable information about the question.

This has been debated quite a few times. The "common knowledge" answer seems to be a 1-roll system.
Assume 15% crit rate, lvl 63 mob (83% hit), +5% to hit

random 1-100
1-15 = crit
16-88 = hit (83+5)
89-100 = resist

Crit is capped by hit (as shown by elemental mastery and in the melee system itself), so you cannot have more crit than you have hit.

• bartolimu · 08/02/06, 4:05 PM

Thanks Aghama, that's the post I was taking about, right down to the gratuitous swearing. :)

balorina · 08/02/06, 5:42 PM

Originally Posted by Skav

The mana regeneration is what you can get from buffs, mp/5s on items or enchants, I consider (which is mostly true) that a warlock spends most of the time inside the 5 second rule interval.

Originally Posted by balorina

Your simulator is coming up with numbers that are too low. It is in how you calculate average shadow bolt damage and filtering down from there.

Care to explain ?

Btw, I assumed that hit and crit are independant rolls, ie crits can be resisted and don't work like melee. I couldn't find any reliable information about the question.

Post your formula and I'll correct it for you. Even if you you do two rolls, your formulas are coming up with numbers that are lower than they should be.

Skav · 08/03/06, 3:10 PM

Originally Posted by balorina

The "common knowledge" answer seems to be a 1-roll system.

Never trust the common knowledge

Lvl 44 warlock vs lvl 53 scorpion
15% theoretical crits on searing pain with talents, + some from int and base (the formula for levels other than 60 is not known)
Theoretical hit chance : 17%

200 searing pains
22 hits (11%)
3 crits (14%)

So a 1-roll system would be very unlikely.

My formula :
sb = (482 + 538) / 2) + (shadow_bonus * 0.857)
sb = sb * (1 + bonus), bonus being the multiplication of various talent +%
sb = sb * hit chance
sb = sb * (1 - partial resists)
sb = sb * (1 + 1.64 * crit / 100), 1.64 = 1 (for ruin damage bonus) + 0.64 (improved shadowbolt with 80% efficiency http://www.skav.org/theorycraft/warlock_impsb.php )
sb = sb * 1.1 (curse of shadow, if present)

Don't think I forgot anything

TheOnly · 08/07/06, 5:11 PM

Originally Posted by Mosh

Originally Posted by Moleva

I don't see how it's any different.

The longer the uptime the greater benefit it provides to all warlocks.

For my simple 2 warlock system, say that one warlock has a 10% crit rate and the other has a 15% crit rate. The uptime for the debuff is 65.8% of the time. If there was only the 10 crit warlock, the uptime would be 34.39%, and for only the 15 crit warlock, the uptime 47.8%. Now obviously the 10 crit warlock gets a greater benefit from having the 15 crit warlock there than vice versa, but the 15 crit warlock gets a fair increase to his dps none the less. And this is at a pretty substantial difference in gear. With both at 15 crit, the uptime is 72.75%. That's a huge gain for both warlocks.

This is wrong.

Uptime is 1-(1-avgcrit)^4, avgcrit being the average critrate of all Warlocks in the raid. In the 10/15% case assuming both Warlocks have an equal manapool and the same spec with regards to mana usage, the uptime will be 41.4%

Faulty assumptions.

Your math assumes that every warlock is chain casting, and gets 4 bolts off in the time duration of the debuff. In a real situation, because there are curses/dots/life tap also occuring, the average duration between bolt casts is longer, and only 3 or so will land during the duration of the debuff. This makes the average duration of the debuff increase slightly as you add warlocks, even if they all have the same crit%.

The reason is simple. If I can cast 4 bolts during the debuff time and I'm by myself in a raid, then I use every charge possible no matter what. Adding a second, identical, warlock consumes the charges twice as fast and also puts them up twice as fast.

However, if I can only cast 3 bolts in the time frame on average and I am alone as the sole warlock, there is an equilibrium average time that the debuff is up, but some charges go to waste. Add a second, equivalent warlock and the likelyhood of a charge going to waste drastically drops yet the rate that the debuff is put up is doubled. Since there was on average 1 wasted debuff charge, adding a second warlock only consumes the debuffs 50% faster but puts them up twice as fast. Overlapping effects diminish this benefit, but no matter how you look at it more (identical) warlocks in the raid means that debuff is up more often simply because a single warlock can't consume all his own charges effectively due to the inability to chain cast.

Zoner · 08/07/06, 6:22 PM

Originally Posted by Skav

Originally Posted by balorina

The "common knowledge" answer seems to be a 1-roll system.

Never trust the common knowledge

Lvl 44 warlock vs lvl 53 scorpion
15% theoretical crits on searing pain with talents, + some from int and base (the formula for levels other than 60 is not known)
Theoretical hit chance : 17%

200 searing pains
22 hits (11%)
3 crits (14%)

So a 1-roll system would be very unlikely.

With that data the margin of error for 22 hits is 5.69% and 3 crits is 2.214%. And even then it can still be off by more than that though its pretty unlikely.

Skav · 08/07/06, 6:34 PM

My point was : with theoretical 17% hit and 15%+ crit, you should see mostly crits with a 1-roll system.

Zoner · 08/07/06, 7:25 PM

Originally Posted by Skav

My point was : with theoretical 17% hit and 15%+ crit, you should see mostly crits with a 1-roll system.

Not if it caps at a base 75% miss rate