Jump to content


Photo

Demonic Pact and rDPS vs Personal DPS


  • Please log in to reply
169 replies to this topic

#1 fallenman

fallenman

    Piston Honda

  • Members
  • 248 posts

Posted 06 February 2009 - 08:03 PM

I'd like to open up an in depth discussion about the merits of demonic pact.

The basic comparison I'm looking at is determining the loss of personal DPS by using a spec that has DP, and weighing that against the possible increase in rDPS. Now, there are a lot of variables to consider in this comparison, including the following:

-Gear level of the warlock. Based upon scaling of a Haunt/Ruin spec vs a DP spec, the more gear, the bigger the loss in personal DPS. (This is based off of SimCraft simulations of scaling) So, how much personal DPS is lost.
-Elemental Shaman presence. If there is an elemental shaman in the raid, then the benefit to the caster DPS will not be as great since you are only providing spell power equivalent to your DP buff minus the 280 from totem of wrath. So for example, if your DP buff provides 350 spell power, that's a net gain of 70 spell power to each caster in the raid.
-DP uptime. This will vary with gear (crit rate), and would need to be calculated, and ultimately factored in.
-Individual caster DPS increase. This would be the biggest factor. Basically, what we would need to determine is what the personal dps increase would be for each class. For example: An arcane mage that does 6000dps would see X dps increase from the additional spell power from DP. A shadow priest that does 6000dps would see X dps increase from Y spell power increase. Etc, and so on.
-Raid composition. How many of each of the above casters are in the raid.

When I have the time, I'll be trying to use SimCraft to try and come up with answers to some of these, but I think this is certainly worth an in depth discussion from all who have input, especially the fantastic theorycrafters who are here. :)

#2 KharzaXo

KharzaXo

    Glass Joe

  • Members
  • 19 posts

Posted 06 February 2009 - 08:32 PM

As a note to using simcraft to answer these questions..

There are still some flaws I believe in the simcraft logic to helping determine some of these answers. I ran one in the demonology thread and pointed out some of the problems I saw. I don't think it takes into account the proper melee buffs on the felguard around crit bonuses. It also doesn't directly have a demonic pact up time statistic to monitor this. I also know there where issues being able to properly account for demonic empathy up time because it didn't seem to only use cleave crits, but rather all crits. With that said the current demonic empathy numbers it puts out, but be a better way to gauge demonic pact up time in simcraft's current form.

I know when I was running the sims for my gear level and monitoring real world raid performance my DP up times came pretty close to the DE up times that simcraft would report. I think some of these issues are why you have seen a decent amount of us demon guys move towards real world parsing to be able to answer some of these questions. Maybe, however, we can as a community provide some accurate information and get some changes to simcraft to help us model this though.

Here is the link to the post I made running various meta/ruin specs through the simcraft project with 100,000 iterations

#3 Trickykid

Trickykid

    Don Flamenco

  • Members
  • 478 posts

Posted 06 February 2009 - 09:10 PM

Pretty simple, once you have the right info:

Average Added Spell Power
Without elemental shaman (only 144 spell power from totem):
DP_uptime*(lock_spellpower-1440)/10.

With elemental shaman (280 spell power):
DP_uptime*(lock_spellpower-2800)/10.

rDPS gain from +1 spell power
Then we need to know how much the raid benefits from the additional stats.

Using SampleOutput - simulationcraft - Google Code for DPS per +1 spell power by class, we get:

Druid 58/0/13 SF     1.23
Druid 58/0/13 W      1.13
Druid 58/0/13 W FF   1.09
Mage 0/53/18         1.45
Mage 18/0/53         1.34
Mage 20/51/0         1.45
Mage 52/19/0         1.27
Mage 57/3/11         1.33
Priest 14/0/57       1.33
Priest 14/0/56 noSWD 1.36
Priest 38/28/5       0.65
Shaman 16/55/0       0.62
Shaman 57/14/0       1.15
Lock 0/31/40         1.06
Lock 0/41/30         1.03
Lock 0/53/18         1.05
Lock 2/13/56         1.13
Lock 55/0/16         1.5

Multiply the number of each of those class/specs by the number above and then sum those numbers. (For example if you have one 58/0/13 SF druid, two 57/3/11 mages, one enhancement shaman and 2 55/0/16 locks, you would get 1*1.23+2*1.33+1*0.62+2*1.50 = 7.51).

Calculate!
That's the raid DPS gained per point of spell power. Multiply that number by the calculated average spell power added to the raid from above. (For example, you have 90% uptime, and 2800 spell power and no ele shaman, you get 0.9*(2800-1440)/10 = 122.4 added spell power. 122*7.51 = 919 added rDPS).

EDIT: Correct FT totem buff used.

#4 fallenman

fallenman

    Piston Honda

  • Members
  • 248 posts

Posted 06 February 2009 - 10:33 PM

Wouldn't flametongue be providing 144 spell power? So, using the numbers:

0.9*(2800-1440)/10 = 122.4. 122.4 * 7.51 = 919 rDPS?

#5 Dramah

Dramah

    Glass Joe

  • Members
  • 11 posts

Posted 06 February 2009 - 10:39 PM

Wouldn't flametongue be providing 144 spell power? So, using the numbers:

0.9*(2800-1440)/10 = 122.4. 122.4 * 7.51 = 919 rDPS?



This.


AND, Other things to consider:

Healing - Even a "small" increase of 70sp really helps HoT's, AoE heals and of course all direct heals. We can't just be narrow minded and look at dps only.

Shaman - There is also the possible increase in the shaman's dps w/ his new found freedom to drop Fire Elemental Totem on cd. I have absolutely no idea what an Fire Elemental hits for anymore, so don't hate me if this is way out in left field.

#6 Trickykid

Trickykid

    Don Flamenco

  • Members
  • 478 posts

Posted 06 February 2009 - 10:57 PM

Wouldn't flametongue be providing 144 spell power? So, using the numbers:

0.9*(2800-1440)/10 = 122.4. 122.4 * 7.51 = 919 rDPS?


Again, uptime, spell power and #casters were pure example. Inputs needed for ilvl213+ estimates:
-Lock spell power with a DP spec and ilvl213+ gear
-Parses from DP locks for actual uptimes

The variable for number of casters is pretty huge, obviously. Also I didn't see Ret pallies on the list from simcraft, and they may gain from this as well. Including a "replacement" fire totem is another good point even if it's the fire ball one.

#7 Morrigan

Morrigan

    Von Kaiser

  • Members
  • 83 posts

Posted 06 February 2009 - 11:27 PM

Unfortunately, DP still doesn't take the spell power from the spirit part of Fel Armor into account, and I doubt that'll be fixed before 3.1 goes live. This makes it hardly possible to reach values over 2800 sp (without spirit contribution) with the currently available gear, but the talent has great potential if they fix this and Ulduar gear becomes available.

#8 PigTrickster

PigTrickster

    Glass Joe

  • Members
  • 3 posts

Posted 07 February 2009 - 12:35 AM

I've been working on parts of this.

One thing that in particular bothered me was the Uptime of Demonic Pact.
So I took a Naxx25 log and scanned three boss fights. The DP Uptime was
Loetheb: 87%
Noth 90%
Razuvious: 89.8%

I've since parsed two more tests and have not been able to break 92%.
So I'm using 90% in my calculations.

My raid buffed pre-engagement spell power is running about 2850.
But when trinkets and demonic pact go off I've witnessed up to 4380 Spell Power.
More typical in fight Spell Power is about 3100 to 3300.

Since we don't have an elemental shaman this is a no-brainer.
But I'd really like to get one and then determine if it's better for the
raid if I'm spec'd Demon or Affliction.

#9 Vindra

Vindra

    Glass Joe

  • Members
  • 5 posts

Posted 07 February 2009 - 07:36 AM

I've been working on parts of this.

One thing that in particular bothered me was the Uptime of Demonic Pact.
So I took a Naxx25 log and scanned three boss fights. The DP Uptime was
Loetheb: 87%
Noth 90%
Razuvious: 89.8%

I've since parsed two more tests and have not been able to break 92%.
So I'm using 90% in my calculations.

My raid buffed pre-engagement spell power is running about 2850.
But when trinkets and demonic pact go off I've witnessed up to 4380 Spell Power.
More typical in fight Spell Power is about 3100 to 3300.

Since we don't have an elemental shaman this is a no-brainer.
But I'd really like to get one and then determine if it's better for the
raid if I'm spec'd Demon or Affliction.


What pet did you use? I experimented with a DP build that utilized a imp over a FG for a few reasons:
1. The imp with talents attacks every 2 seconds and with a 12 second DP buff he only needs to crit at about a 40% rate to have a 99 uptime.
2. Master Demo buff is fire damage and crit for imp and master. I used a rotation similar to what a 0/31/40 or 0/40/31 build would.
3. Demonic empower on a FG is a very tiny DPS increase. DE on a imp is crit, very useful to pop at the start of a fight to get DP up fast.
4. Demonic Empathy is up almost 100%. Felgaurd cleave is the only special to proc it and that has a CD while the imps firebolt does not.

I think a FG/shadowbolt spec would have slightly higher personal DPS numbers than a Imp/fire both providing DP. I opted to maximize my raid contribution over personal DPS. I could just spec affliction and have a shaman drop flamgetounge and probably net the same result DPS wise, not when counting healers. Assuming when they fix the spirit to spell damage on FA and nothing else changes a DP lock will be a requirement for a raid.

#10 Trickykid

Trickykid

    Don Flamenco

  • Members
  • 478 posts

Posted 07 February 2009 - 08:21 AM

1. The imp with talents attacks every 2 seconds and with a 12 second DP buff he only needs to crit at about a 40% rate to have a 99 uptime.


1-(1-.4)^6 = 95% uptime. How are you calculating 99%? The FG should be attacking more often than 2 seconds anyway, no? His base swing is 2s, which can be hasted, and he also has cleaves. He also has substantially more DPS than the imp. It's not worth sacrificing total DPS just to buff others more, point of this thread is to figure out how much is added through DP. With the Fel Armor bug, it's probably still shy of making up for the loss from going demo over affliction. Adding to that by downgrading to an imp (which is anywhere from 500-800 DPS lost depending on the parses you look at), is not a good idea for personal DPS nor raid DPS.

#11 Feefes

Feefes

    Glass Joe

  • Members
  • 6 posts

Posted 07 February 2009 - 09:44 AM

Unfortunately, DP still doesn't take the spell power from the spirit part of Fel Armor into account, and I doubt that'll be fixed before 3.1 goes live. This makes it hardly possible to reach values over 2800 sp (without spirit contribution) with the currently available gear, but the talent has great potential if they fix this and Ulduar gear becomes available.


It is certainly possible to exceed 2,800 spellpower before spirit. I sit at approximately 3150 spellpower without a totem but fully-buffed otherwise, approximately 275 of which comes from Spirit. I realise that this is a fairly slight increase over ToW, but it's not so impossible as you claim.

#12 dedmonwakeen

dedmonwakeen

    Bald Bull

  • Members
  • 1302 posts

Posted 07 February 2009 - 03:59 PM

As a note to using simcraft to answer these questions..

There are still some flaws I believe in the simcraft logic to helping determine some of these answers. I ran one in the demonology thread and pointed out some of the problems I saw. I don't think it takes into account the proper melee buffs on the felguard around crit bonuses. It also doesn't directly have a demonic pact up time statistic to monitor this. I also know there where issues being able to properly account for demonic empathy up time because it didn't seem to only use cleave crits, but rather all crits. With that said the current demonic empathy numbers it puts out, but be a better way to gauge demonic pact up time in simcraft's current form.

I know when I was running the sims for my gear level and monitoring real world raid performance my DP up times came pretty close to the DE up times that simcraft would report. I think some of these issues are why you have seen a decent amount of us demon guys move towards real world parsing to be able to answer some of these questions. Maybe, however, we can as a community provide some accurate information and get some changes to simcraft to help us model this though.

Here is the link to the post I made running various meta/ruin specs through the simcraft project with 100,000 iterations


The guys taking care of the Warlock module have addressed nearly all of these concerns. I say "nearly" because I think we still need to add the demonic pact uptime. I've just been so busy with family/work/etc issues I haven't put out a new release. One way or another I'll get it done this weekend.......

#13 Vindra

Vindra

    Glass Joe

  • Members
  • 5 posts

Posted 07 February 2009 - 04:18 PM

1-(1-.4)^6 = 95% uptime. How are you calculating 99%? The FG should be attacking more often than 2 seconds anyway, no? His base swing is 2s, which can be hasted, and he also has cleaves. He also has substantially more DPS than the imp. It's not worth sacrificing total DPS just to buff others more, point of this thread is to figure out how much is added through DP. With the Fel Armor bug, it's probably still shy of making up for the loss from going demo over affliction. Adding to that by downgrading to an imp (which is anywhere from 500-800 DPS lost depending on the parses you look at), is not a good idea for personal DPS nor raid DPS.


Demonic Empowerment on CD is an extra 20% crit for 30 seconds every minute. When I tried this build my imp had about 43-44% crit vs a mob with mage and boomkin present before DE. I had parses with 98.9% uptime, I'll see if I can get one to post.

As for Imp vs Felgaurd I would propose that the loss in pet damage is made up for in synergy. The FG MD buff is 5% damage done and 5% damage reduction. The imp MD is 5% fire crit and 5% fire damage done. Crit isn't a monster DPS inc stat but is better than 5% damage reduction. A DP build can't get deep enough in destro to pick up SnF but can get moltencore and ruin. Just like a 0/41/30 build I used Agony to proc MC and immo/incin(corruption when forced to move) Also a imp can keep demonic empathy up 100%. DEmpathy is a 15 second buff, longer than DP, and a FG can only proc it once every 6 seconds with cleave. I give up ISB(10%) and FG MD(5%dam and % reduc) + FGs DPS for Moltencore(10%)Imp MD 5% fire 5% fire crit, higher DP uptime and higher DEmpathy uptime throught the imp DEmpower, and imp DPS.

One thing I think that may be overlooked is the fact that the imp once capped can't miss. He will still suffer from partial resists like all spells do but the Felgaurd will get dodged plus glancing blows. Felgaurds will have less crit since they have none on their MD bonus like an imp, none on their Demonic Empower, and cannot benefit from a mage's 10%.

#14 Trickykid

Trickykid

    Don Flamenco

  • Members
  • 478 posts

Posted 07 February 2009 - 07:44 PM

As for Imp vs Felgaurd I would propose that the loss in pet damage is made up for in synergy. The FG MD buff is 5% damage done and 5% damage reduction. The imp MD is 5% fire crit and 5% fire damage done. Crit isn't a monster DPS inc stat but is better than 5% damage reduction. A DP build can't get deep enough in destro to pick up SnF but can get moltencore and ruin. Just like a 0/41/30 build I used Agony to proc MC and immo/incin(corruption when forced to move) Also a imp can keep demonic empathy up 100%. DEmpathy is a 15 second buff, longer than DP, and a FG can only proc it once every 6 seconds with cleave. I give up ISB(10%) and FG MD(5%dam and % reduc) + FGs DPS for Moltencore(10%)Imp MD 5% fire 5% fire crit, higher DP uptime and higher DEmpathy uptime throught the imp DEmpower, and imp DPS.

One thing I think that may be overlooked is the fact that the imp once capped can't miss. He will still suffer from partial resists like all spells do but the Felgaurd will get dodged plus glancing blows. Felgaurds will have less crit since they have none on their MD bonus like an imp, none on their Demonic Empower, and cannot benefit from a mage's 10%.


Imp requires you to use incinerate which will be worse than SB without emberstorm. It also only affects immolate and incinerate compared to 5% all affecting CoA/Corr/Immo/filler. The other points you make are valid but don't add up if you check parses/math/etc. I'd check out the sims or spreadsheet math on that for your gear, but for me, SB spam comes out ahead. Haven't seen any parses or data that show the lost DPS from FG->imp is regained anywhere.

#15 ExinferisD

ExinferisD

    Glass Joe

  • Members
  • 4 posts

Posted 07 February 2009 - 11:02 PM

I'm not very good with math when it comes to theory crafting but I'll go ahead and post last weeks WWS as me being Meta/Ruin. I'm estimating its a ~500 dps drop on fights like Patchwerk but can be roughly the same or even more on fights that are hectic / involve aoeing. This WWS does contain an Elemental Shaman- he is trying to join our raid force. If he ends up getting a raid spot I will likely go back to Affliction.

Please feel free to use this WWS if you are trying to calculate DP up-time, dps gains for raid vs dps loss for warlock, ect. Also keep in mind that I'm an Alchemist with increases flask effects and I pop Blood Fury on cooldown.

Naxx: Wow Web Stats

Arch (you'll have to add Felguard + my parse yourself):

Wow Web Stats
+
Wow Web Stats

Our Maly WWS seems very bugged and I'll be able to post a 3D after Monday.


I'm very curious if Blizzard intends to buff Totem of Wrath or if they will nerf Demonic Pact. When the fix to Fel Armor + DP goes live and Ulduar gear comes out, DP is going to start blowing ToW out of the water.

#16 PigTrickster

PigTrickster

    Glass Joe

  • Members
  • 3 posts

Posted 09 February 2009 - 05:29 PM

I had originally thought about using an Imp but the talent point required to support
the Imp properly just don't create a viable spec (IMO). So I was using a FG. His dps
contributions are very pleasant and the best consistent pet dps that I can get. Though
the Doomguard clearly outclasses him, but that's only once and hour.

Starting with simple numbers and ignoring the loss of spell power due to spirit.
I calculate that for my raid makeup I need 3110 Spell Power with DP up 90% of the time
to be equal to Totem of Wrath.

Increasing my SP to 3300 I'm add a net 159 DPS to my raid over a ToW.
I'm fairly certain that my personal DPS loss by not being affliction is significantly
greater than that.

My conclusion:
If you do not have an Elemental Shaman and you have a spare warlock then
definitely have one of them spec Demonologist. But this is dependent on raid
composition and the number of casters.

If you do have an Elemental Shaman then go affliction.
At least until the spirit bug is fixed in Fel Armor.

Simple Spreadsheet: https://spreadsheets...vnEPW0TTQ&hl=en

Caveat - I won't claim to be precise on all of these numbers, but the napkin math indicates
accuracy is sufficient for this case, not precision.

#17 Zakalwe

Zakalwe

    The Chairmaker

  • Members
  • 1466 posts

Posted 09 February 2009 - 07:21 PM

As a note to using simcraft to answer these questions..

There are still some flaws I believe in the simcraft logic to helping determine some of these answers. I ran one in the demonology thread and pointed out some of the problems I saw. I don't think it takes into account the proper melee buffs on the felguard around crit bonuses. It also doesn't directly have a demonic pact up time statistic to monitor this. I also know there where issues being able to properly account for demonic empathy up time because it didn't seem to only use cleave crits, but rather all crits. With that said the current demonic empathy numbers it puts out, but be a better way to gauge demonic pact up time in simcraft's current form.

I know when I was running the sims for my gear level and monitoring real world raid performance my DP up times came pretty close to the DE up times that simcraft would report. I think some of these issues are why you have seen a decent amount of us demon guys move towards real world parsing to be able to answer some of these questions. Maybe, however, we can as a community provide some accurate information and get some changes to simcraft to help us model this though.

Here is the link to the post I made running various meta/ruin specs through the simcraft project with 100,000 iterations

This is true - there are two issues in the currently available release of Simcraft: Demonic Pact uptime is not being tracked/reported properly, and Demonic Empathy is being erroneously triggered by melee swings. The DE bug is already fixed in SVN, and I'll make sure the DP reporting issue is fixed before the next release.

#18 fallenman

fallenman

    Piston Honda

  • Members
  • 248 posts

Posted 11 February 2009 - 03:11 PM

In my current gear, I had a DP uptime of around 90% when I was a 0/56/15 spec last week. My gear is mostly ilvl 213+. This was with a 24% cleave crit rate, and 29% swing crit rate.

#19 dahaliana

dahaliana

    Glass Joe

  • Members
  • 4 posts

Posted 12 February 2009 - 08:08 PM

I spec'd Demonology for a 20 man Naxxramas run yesterday and had decent results with it - I'm usually affliction.

Here is something I don't think was mentioned:
If you have an elemental shaman and a paladin that brings the 3% crit buff, then your Demonic Pact allows the shaman to use a Searing Totem. You can effectively add that totem's DPS to your rDPS increase.

It seemed to me that full Demonology was about 1000dps behind Affliction with a Doomguard (for me). I got 5300 on Patchwerk vs. 6300 the previous kill as Affliction with a Doomguard. The gap would be considerably smaller if it weren't for the Doomguard putting out such high numbers.

So I am trying to work out where the breakpoint is. If I can show Demonology to be a rDPS increase of 1100 I will spec it more often for sure, but the useful break may be lower than that in many cases. Affliction numbers go through the roof with a Doomguard.

Anyone have any idea on the DPS put out by a Searing Totem to work that into the calculation?

#20 zerdell

zerdell

    Von Kaiser

  • Members
  • 45 posts

Posted 13 February 2009 - 02:59 AM

Here is something I don't think was mentioned:
If you have an elemental shaman and a paladin that brings the 3% crit buff, then your Demonic Pact allows the shaman to use a Searing Totem. You can effectively add that totem's DPS to your rDPS increase.


Demonic Pact won't give you 100% uptime on the spellpower buff. I haven't run any number,s but I suspect that if you get say 80% uptime from it, the extra raid damage from 20% with ele totem over non ele totem would out damage searing totem.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users