SimulationCraft for Warlocks (4.0.3a numbers at 80 and 85) - Page 80

Scrufola · 04/17/09, 7:36 AM

Could you add an option like the debuff=1 of haunt for shadow bolt? We could then calculate the sacrifice for a destro lock if he has to keep up and skill for the 5% crit debuff.

exog · 04/17/09, 8:06 AM

Originally Posted by Aeiri

Are those numbers right? I mean I can understand affliction being up at the top with its scaling with all the new gear...but only a ~300dps different from using a doomguard and a succubus? Am I missing something here...

Wasn't doomguard dps nerfed along with hp and summoning time?

Tinava · 04/17/09, 8:28 AM

Originally Posted by exog

Wasn't doomguard dps nerfed along with hp and summoning time?

All pets were nerfed a bit, as I recall. Including the DG.

Asmodaeus · 04/17/09, 8:47 AM

Before the nerf to Conflagrate, it was shown here that without the 4p t7 bonus, the highest dps spec for Deep Destro is one with Improved Soul Leech, and using Glyph of Incinerate instead of Glyph of Life Tap. Since you've updated the OP by replacing tier 7 with tier 8, could you please test this now and update the OP if it's still true?

téhanu · 04/17/09, 8:59 AM

Originally Posted by Asmodaeus

Before the nerf to Conflagrate, it was shown here that without the 4p t7 bonus.... using Glyph of Incinerate instead of Glyph of Life Tap.

So you mean in any case, when not getting the full benefit from life tap (no 4p t7), we should rather glyph incinerate than LT?

...this'll take the lt out of the early rotation too.

Asmodaeus · 04/17/09, 9:55 AM

Originally Posted by téhanu

So you mean in any case, when not getting the full benefit from life tap (no 4p t7), we should rather glyph incinerate than LT?

...this'll take the lt out of the early rotation too.

It seems to depend on gear and being specced into ISL. Having more spirit and 4t7 favors the Glyph of Life Tap. Having more intellect and 4t8 seems to favor ISL and Glyph of Incinerate.

Drison · 04/17/09, 11:54 AM

Originally Posted by Tinava

All pets were nerfed a bit, as I recall. Including the DG.

Only the DG was nerfed due to its cooldown being reduced to 30mins and the voidwalker had the amount of health it gains from us reduced to 10%.

Scrufola · 04/17/09, 2:06 PM

With the nerfed Conflagrate, the Incinerate glyph would be better than the Immolate glyph for 03/13/55.

Originally Posted by Zakalwe

Detailed damage breakdowns:

Player=Warlock_T8_03_13_55  DPS=6780.7 (Error=+/-1.8 Range=+/-565)  DPR=22.0  RPS=308.8/262.5  (mana)
  Core Stats:  strength=54  agility=55  stamina=1112  intellect=978  spirit=507  health=14465  mana=18684
  Spell Stats:  power=2348  hit=11.1%  crit=20.0%  penetration=0  haste=13.8%  mp5=0
  Attack Stats:  power=44  hit=8.9%  crit=16.2%  expertise=0.0  penetration=0.0  haste=13.8%
  Defense Stats:  armor=110
  Actions:
    chaos_bolt            Count= 20.4|14.8sec  DPE=10301|10%  DPET= 7353  DPR=  42.3  Miss=0.0%  Hit=6801  Crit=14214|16775|47.2%
    conflagrate           Count= 27.7|10.9sec  DPE=10815|15%  DPET= 8070  DPR=  19.4  Miss=0.0%  Hit=6913  Crit=14436|17095|51.9%
    curse_of_doom         Count=  4.0|61.5sec  DPE=16609| 3%  DPET=12138  DPR=  30.5  Miss=0.0%  TickCount=4  Tick=16621
    immolate              Count= 19.0|15.9sec  DPE=13119|12%  DPET=11705  DPR=  22.2  Miss=0.0%  Hit=2490  Crit= 5200| 6150|46.1%  TickCount=93  Tick=1919
    incinerate            Count=120.5| 2.4sec  DPE= 8310|49%  DPET= 5358  DPR=  17.1  Miss=0.0%  Hit=5310  Crit=11098|13421|51.8%
   imp
    fire_bolt             Count=161.1| 1.9sec  DPE= 1262|10%  DPET=  681  DPR=   0.0  Miss=2.9%  Hit=1192  Crit= 1788| 2051|17.5%

The damage gained from the Immolate glyph is:

((93 * 1919 + 27.7 * 10815) / 1.1 ) * 0.1 = 43458

And the Incinerate glyph would increase the damage by:

120.5 * 8310 * 0.05 = 50067

And wowhead changed the talent planer link URL. Simcraft does not understand the new URL, you have to replace the # with a =.

Sedona · 04/17/09, 2:52 PM

Originally Posted by Zakalwe

I've updated the OP to reflect new simcraft standards now that 3.0 mechanics are irrelevant: T8 is Ulduar gear (best known set), T7 is pre-Ulduar gear (best set). I'm aware that my T8 set is pretty outdated, so feel free to post better suggestions.

Is chardev updated with the latest known Ulduar gear now?

I posted this a while back before 3.1 went live. I am sure it was a bit soon. And debating best in slot gear is still something that will be changing. I'll put my two cents in anyhow.

I personally feel for the T8 set that needs extra hit should be picking up Living Flame or just keeping Dying Curse instead of Flare of the Heavens. These trinkets are still useful as they were pre-patch to free up stats on other gear. And Crit rating still remains to be one of our weakest stats.

Switching these trinkets maybe your ultimate goal once we have a comprehensive list of gear.

I also know that there was a debate between the MH/OH and the Staff Endless Winter with the new Staff Enchants.

Kudos once again for your hard work.

nenad · 04/18/09, 12:12 AM

I did some tests on Target Dummy, and initially it seemed like there are significant differences, but after some double-checking of my gear setup, I managed to get closer numbers. But comparing was not so easy because I had to look at simulator data in one format (output TXT mostly), and at test data in other format (wowmeteronline or WWS). And while it is not problem for checking averages, I was never sure if averages on both sides were calculated same way. Output file for example does not have total damage per any spell, nor DPS per spell. But even more differences are in how they calculate/show spell or effect uptimes.

After losing about hour in manual comparing few tests, I decided to add some code to simulator to be able to generate standard WoW compatible log file if "log=1" option is on. Since I saw simulator code first time few days ago, I lost bit more that I expected (another several hours) doing that, so it would be probably faster if i just continued manually, but ... ;p

Anyway, example of end result is something like:

Original WoW log from TargetDummy, on WMO

Simulated log, also on WMO

So far I only tested 0_41_30, and some conclusion would be that simulator is fairly close. For some reason it shows slightly higher average damages than real test, slightly higher crits and lower misses - which ends up in higher DPS shown on simulator. When you include my far from perfect weaving on real test, you get 3370 DPS on simulator vs 2870 DPS on test (again, no external buffs, solo , on boss dummy). In addition, some uptimes are higher (can be seen in "Buffs and Debuffs on WMO, and I dont mean Decimate one since it is artifically high on TargetDummy because it trigger from start ... one that is higher is Demonic Soul, but may be due to how WMO calculate it).

But since I spent more time making it easier to compare than testing, few more tests should give more precise result, and I plan also to test and compare Haunt build.

On a side note, it would be nice if option to generate standard log files could be added to regular version of simulator, since it greatly simplify comparing. Beside seeing both data in same way, both tools (WMO&WWS) have good log search/filter options for event comparison. If someone who manage simulator code is interested in that, you can PM me and I can send you what I did so far.

Edit: I tested on WWS too , and it works also:
Warlock_T8 raid on WWS
Raid_T8.simcraft on WWS

Knasen · 04/18/09, 9:26 AM

Anyone that have done some serious math on Meta vs Immo glyph as DP-build? Some very basic math make them look pretty similar (around 1% more dmg). Those extra 6 seconds can be pretty important when you got hero and/or decimate is there someone which have done the calculation abit more serious?
Simcraft says Immo glyph but my oversimplified math says Meta. Any input is welcome.

nenad · 04/18/09, 9:42 AM

BTW, I redid my previous tests on normal boss dummy (lvl 80), instead of lvl 83 one that I used previously, because miss chances are unrealistic (since in raid setup we have hit buffs). Also I simplified test without using "decimate" part , because that is one which on simulator is always done more perfectly than on dummy.

Results ended up much closer this time (this is still for 0/41/30 @ 1928dmg, 306hit, 448crit, 327haste, 660spi, eff:Sundial+DCurse+T7_4pc):

Real 0/41/30 test on dummy(lvl80) without decimate

Simulated same situation, on WMO

I noticed some things that are still slightly different tho:
1) average incinerate damage is still higher on seimulator
2) pet did not have any miss on simulator, and had ~3.5% misses on test

Second one is not so noticeable, because pet also had slightly higher average damage per hit on real test compared to sim, and when misses are included it ended up almost same.

First one is probably due to Molten Core uptime, which is much higher on simulator (60% vs 20% on test), even if number of shadow dot ticks is almost same (126 on sim, 120 on test). When I did 50k iteration test on sim, MC uptime was still 65%, so maybe I was just unlucky on test dummy - which looks possible since fast math also says MC uptime at 2/3 MC should be higher.

Edit 1: I retested on lvl 83 dummy, with Decimation, but used more +hit gear (for 0/41/30 @ 1902dmg, 426hit, 354crit, 263haste, 667spi, eff:Sundial+DCurse+T7_2pc). Results are now much closer:

Test on lvl 83 dummy ~ 3250 dps

Simulation of same case, ~3450 dps

while simulator still have slightly higher average damages per spell, it can be explained by higher LT_glyph uptime (I didnt LT during decimate). But more impoertant here is that main thing I changed from previous tests is "hit" - and in real tests it seems that negative impact from not enough hit is higher than on simulator - which can be expected, since unlike me, simulator can immediatelly notice if it missed ;p

Edit 2: Did also affliction build test, on lvl 83 dummy (for 53/1/17, same gear as above in Edit1):

Test on lvl 83 dummy ~ 4100dps

Simulation for same case, ~3900dps

Here I had opposite situation than at demo - real test showed higher numbers than simulation. BUT due to crowded dummies, several people were testing during my test, so it is possible that some additional debuff was on dummy. Even if i couldnt find such debuff in log, it could have been put before log started. As it is now, average spell damage for spells like SB or Corruption is 16% higher on test. Some 8% can be explained by Eradictions 12% being present 65% longer, but will need test without anyone else on dummy to be sure for reason of remaining 8%.

On the other side, crit rate for SB on simulation (~30%) is significantly higher than on test (14%), and that can not be explained by debuffs on dummy. When run 50k iterations, crit on SB was 25%. One reason could be much higher uptime for Demonic Soul on sim, but would have explained only 2-3% of difference - maybe was just unlucky on SB crits during test.

BTW, seems that some WMO links from above expired - maybe because I was posting them as private logs, or maybe those links to players are not persistent. I refreshed last link, this time with link to fight not to player. Will refresh others when next time I do tests.

Soygen · 04/18/09, 10:20 AM

Quick question on the simulation results. What does the "_DG" at the end of the top-performing simulations mean?

jeanvaljean · 04/18/09, 10:31 AM

Originally Posted by Soygen

Quick question on the simulation results. What does the "_DG" at the end of the top-performing simulations mean?

It means that a doomguard is used as your pet, meaning that it's only possible to do for 15 mins every half hour.

Xzael · 04/18/09, 12:01 PM

warlock=Warlock_T8_53_00_18
level=80
talents=http://talent.mmo-champion.com/?warl...00000000000000
actions=flask,type=frost_wyrm/food,type=tender_shoveltusk_steak/spell_stone/fel_armor/summon_pet,succubus
actions+=/haunt,debuff=1/corruption/curse_of_agony/unstable_affliction/haunt
actions+=/drain_soul,health_percentage<=25,interrupt=1/shadow_bolt/life_tap
Warlock_T8_Gear.simcraft
glyph_haunt=1
glyph_curse_of_agony=1
glyph_life_tap=1

I tried looking on the Simulationcraft page for a guide or explanation as to how to read the "actions" line, but I couldn't find one. Could anyone point me to the right reference or help me understand how to read these two lines:

actions+=/haunt,debuff=1/corruption/curse_of_agony/unstable_affliction/haunt
actions+=/drain_soul,health_percentage<=25,interrupt=1/shadow_bolt/life_tap

How does it prioritize spells, or is that based in the Simulationcraft logic, and this is just a list of abilities it will use (I'm assuming it's not since haunt is listed twice)? Also, what does the "debuff" mean in "haunt,debuff=1", any why don't the other abilities have that parameter?