Archived

This topic is now archived and is closed to further replies.

Starfox

SimulationCraft - S(t)imulating your inner rage

96 posts in this topic

That's the way i implemented it, fine

What about 3.1.0 now with weapon mastery switched to arms?

This would likely be new standard fury in 3.1 then: Talent Calculator - World of Warcraft

Also, I'll need to update that gearset too, it was pre-3.1 changes, and pretty bad after the imp zerk stance change.

Share this post


Link to post
Share on other sites

I took a quick glace over your coding, and while only having 1.5 years of Java learning under my belt, I understand the general ideas you are going for.

One thing I noticed that concerned me is your No-Abilities-During-Bladestorm solution. A 6 sec GCD would solve it, EXCEPT for HS/Cleave in actuality, but I'm not sure exactly how you've programmed those abilities, and if Heroic Strike is actually disabled during Bladestorm like it should be.

Otherwise, it looks great. Do you have a UI that we could start plugging numbers into and test results yet?

Share this post


Link to post
Share on other sites
I took a quick glace over your coding, and while only having 1.5 years of Java learning under my belt, I understand the general ideas you are going for.

One thing I noticed that concerned me is your No-Abilities-During-Bladestorm solution. A 6 sec GCD would solve it, EXCEPT for HS/Cleave in actuality, but I'm not sure exactly how you've programmed those abilities, and if Heroic Strike is actually disabled during Bladestorm like it should be.

Otherwise, it looks great. Do you have a UI that we could start plugging numbers into and test results yet?

Simulationcraft doesn't have a UI. You just edit the .simcraft files and (on windows) double click on them to run it (if configured)/drag and drop them on top of the simcraft.bat file.

Share this post


Link to post
Share on other sites
It works pretty much like you said, will never use more than 30 rage, will never leave you with less than 10 rage, sooo:

40+ rage: only 30 rage is lost and damage is the same as a 30 rage execute

30-39 rage: damage is the same as a 30 rage execute, but you're left with 10 rage afterwards

<30 rage: damage is the same as a regular execute at whatever rage you're at, and you're left with 10 rage afterwards

As far as I understand for >=30 rage damage is calculated as for 30-execute_cost rage + glyph. I'm basing this on the following patch notes (3.0.8):

Sudden Death: The amount of rage spent on an Execute while Sudden Death is active is now capped at 30 rage. Improved Execute and the Glyph of Execute will still increase the damage done by increasing the amount of rage able to be converted into damage.

Share this post


Link to post
Share on other sites
As far as I understand for >=30 rage damage is calculated as for 30-execute_cost rage + glyph. I'm basing this on the following patch notes (3.0.8):

Right, that's still just the damage of a normal 30 rage execute though. A 30 rage execute with imp execute/glyph will have 30*38 bonus damage, and a 30 rage execute without them will have 15*38 bonus damage. I only meant that the calculation is no different for SD.

Share this post


Link to post
Share on other sites

Fix the weirdness of stats on bladestorm, would be nice if someone with an eye for it could look at the stats and spot if something lookss just wrong :)

Share this post


Link to post
Share on other sites

Fixed a few bugs and did a rerun with the latest version of the sim, can anyone fimiliar with fury/arms look at the numbers and say if they are reasonable or just out of line?

Besides that, what glyphs are now used?

Fury: WW, HS, Exe ?

Arms: MS, Rend, Bladestorm ?

Share this post


Link to post
Share on other sites

Besides that, what glyphs are now used?

Fury: WW, HS, Exe ?

Arms: MS, Rend, Bladestorm ?

Mostly, however Exe glyph is better than Bladestorm

Share this post


Link to post
Share on other sites
Mostly, however Exe glyph is better than Bladestorm

Thx, will change glyphs and redo the run

Share this post


Link to post
Share on other sites

The arms actions you've listed are probably not quite right. Here's what I'm thinking:

actions=flask,type=endless_rage/food,type=dragonfin_filet

actions+=/stance,choose=battle/auto_attack

actions+=/bloodrage,rage<=85

actions+=/rend

actions+=/execute,sudden_death=1

actions+=/overpower

actions+=/bladestorm

actions+=/heroic_strike,rage>=90

actions+=/mortal_strike

actions+=/slam

Can you try those out? :) I'll try to get & compile your changed version as well to test it myself...

Ervin

Share this post


Link to post
Share on other sites

Ok, I managed to run some tests, here's what I found to be the "best", with the current simulation implementation at least:

actions=flask,type=endless_rage/food,type=dragonfin_filet

actions+=/stance,choose=battle/auto_attack

actions+=/bloodrage,rage<=85

actions+=/rend

actions+=/overpower

actions+=/bladestorm

actions+=/mortal_strike

actions+=/execute

actions+=/heroic_strike,rage>=80

actions+=/slam

Looks rather simple priority order. Heroic striking raised the dps only ~1% though, so might as well forget it and concentrate on slamming efficiently. Adding criteria to only execute if have >=30 rage lowered the dps.

Thanks a ton Starfox for implementing the warrior into simcraft - I had started it myself too but got distracted by other issues ;).

Share this post


Link to post
Share on other sites
Ok, I managed to run some tests, here's what I found to be the "best", with the current simulation implementation at least:

Have you tried it with [iTEM]Hearty Rhino[/iTEM] and [iTEM]Elixir of Armor Piercing[/iTEM] instead of the STR/AP buffs?

Edit: OK, kids, this is why posting on no sleep because you're on call and got called all night long is bad. Obviously 45 ArP, no matter how good, isn't going to be better than 180 AP. But, I am still curious about the ArP food over [iTEM]Dragonfin Filet[/iTEM].

Share this post


Link to post
Share on other sites

One things that boggles my mind at the moment is the scaling values for agi and crit for fury/arms

Scale Factors:

  Warrior_T7_18_53_00        agility=0.59  crit_rating=1.18

  Warrior_T7_MS              agility=0.73  crit_rating=0.89

I'll look into that after raiding, pretty weird though :/

Share this post


Link to post
Share on other sites

Overall, I'm not getting the same numbers as Starfox - you're not using the globals etc. file from svn? Makes it slightly difficult to compare the figures I'm getting to ones you have...

Also, the arms spec that you used is a bit off I think, a more normal/accepted(?) would be Talent Calculator - World of Warcraft.

Improved execute is seen as giving considerable dps, but not according to the sim. I think the talent is perhaps calculated wrong? When calculating the damage an execute does, shouldn't the 5 rage you get from the talent be added to the "excess rage" that is used to calculate the damage? I did that change, but it didn't make a big impact though. Maybe the talent just is not that good.

Share this post


Link to post
Share on other sites

Ok, I just wasn't using the T8 globals file, silly me. Here's what I get with your original and my "better" arms spec, including the imp.execute change I outlined above. Seems that the heroic strike critting has a bigger impact than imp.exec. I wonder if something is wrong or if that is really how it is.



SimulationCraft for World of Warcraft build 3.1.0 (iterations=1000, max_time=300, optimal_raid=1)


DPS Ranking:

  16443 100.0%  Raid

   5627  34.2%  Warrior_T7_18_53_00

   5439  33.0%  Warrior_T7_MS_Orig

   5398  32.8%  Warrior_T7_MS_2ptImpExec


Player=Warrior_T7_18_53_00  DPS=5627.2 (Error=+/-12.7 Range=+/-695)  DPR=345.2  RPS=16.3/16.4  (rage)

  Core Stats:  strength=687  agility=713  stamina=1263  intellect=41  spirit=71  health=17029  mana=20

  Spell Stats:  power=0  hit=6.3%  crit=21.0%  penetration=0  haste=8.4%  mp5=0

  Attack Stats:  power=3571  hit=5.0%  crit=35.6%  expertise=4.5  penetration=17.9  haste=8.4%

  Defense Stats:  armor=1426

  Actions:

    bloodthirst           Count= 49.6| 5.9sec  DPE= 5243|15%  DPET= 3177  DPR= 174.8  Miss=0.0%  Hit=3204  Crit= 7287|10650|51.5%  Dodge=2.0%

    deep_wounds           Count=  0.0| 0.0sec  DPE=    1|15%  DPET=    1  DPR=   0.0  Miss=-1.$%  TickCount=215  Tick=1215

    execute               Count= 14.0| 3.1sec  DPE= 8207| 7%  DPET= 4974  DPR= 113.9  Miss=0.0%  Hit=4874  Crit=11195|14766|54.3%  Dodge=2.0%

    heroic_strike         Count= 98.3| 2.7sec  DPE= 4010|23%  DPET=    1  DPR= 334.2  Miss=0.0%  Hit=2447  Crit= 5568| 7396|51.7%  Dodge=2.0%

    melee_main_hand       Count= 37.0| 8.2sec  DPE= 2708| 6%  DPET= 1412  DPR=   0.0  Miss=19.0%  Hit=2122  Crit= 4358| 5742|51.9%  Glance=1583|24.1%  Dodge=1.9%

    melee_off_hand        Count=133.9| 2.2sec  DPE= 1655|13%  DPET=  819  DPR=   0.0  Miss=18.8%  Hit=1286  Crit= 2652| 3590|52.3%  Glance= 966|24.0%  Dodge=2.0%

    slam                  Count= 18.7|14.5sec  DPE= 4395| 5%  DPET= 2664  DPR= 293.0  Miss=0.0%  Hit=2689  Crit= 6121| 8243|51.2%  Dodge=1.9%

    whirlwind             Count= 71.0| 8.4sec  DPE= 3514|15%  DPET= 4259  DPR= 281.1  Miss=0.0%  Hit=2125  Crit= 4808| 8054|53.3%  Dodge=2.0%


Player=Warrior_T7_MS_2ptImpExec  DPS=5398.3 (Error=+/-8.9 Range=+/-506)  DPR=424.1  RPS=12.7/12.8  (rage)

  Core Stats:  strength=1103  agility=487  stamina=1314  intellect=41  spirit=71  health=17534  mana=20

  Spell Stats:  power=0  hit=9.8%  crit=20.4%  penetration=0  haste=6.4%  mp5=0

  Attack Stats:  power=3965  hit=7.8%  crit=31.3%  expertise=6.8  penetration=0.0  haste=6.4%

  Defense Stats:  armor=974

  Actions:

    bladestorm            Count=  4.0|91.1sec  DPE=21069| 5%  DPET= 3370  DPR= 842.8  Miss=0.0%  TickCount=28  Tick=1916  CritTick=4146|4253|49.1%

    deep_wounds           Count=  0.0| 0.0sec  DPE=    1|15%  DPET=    1  DPR=   0.0  Miss=-1.$%  TickCount=211  Tick=1178

    execute               Count= 18.0|15.5sec  DPE= 6610| 7%  DPET= 4007  DPR= 133.3  Miss=0.1%  Hit=3916  Crit= 9376|13327|49.4%

    heroic_strike         Count= 23.1|10.8sec  DPE= 3993| 6%  DPET=    1  DPR= 332.8  Miss=0.2%  Hit=2380  Crit= 5701| 5806|48.7%

    melee_main_hand       Count= 92.0| 3.3sec  DPE= 2968|17%  DPET= 1276  DPR=   0.0  Miss=0.2%  Hit=1962  Crit= 4245| 4340|49.4%  Glance=1472|24.0%

    mortal_strike         Count= 43.3| 6.8sec  DPE= 4562|12%  DPET= 2765  DPR= 152.1  Miss=0.2%  Hit=2694  Crit= 6453| 6572|49.8%

    overpower             Count= 46.7| 6.3sec  DPE= 5504|16%  DPET= 4787  DPR=1100.7  Miss=0.2%  Hit=2316  Crit= 5535| 5653|99.1%

    rend                  Count= 14.0|21.7sec  DPE= 7903| 7%  DPET= 4790  DPR= 790.3  Miss=0.1%  TickCount=96  Tick=1152

    slam                  Count= 57.9| 4.2sec  DPE= 4028|14%  DPET= 2442  DPR= 268.5  Miss=0.1%  Hit=2390  Crit= 5725| 5842|49.2%


Player=Warrior_T7_MS_Orig  DPS=5438.6 (Error=+/-8.5 Range=+/-388)  DPR=428.8  RPS=12.7/12.8  (rage)

  Core Stats:  strength=1103  agility=487  stamina=1314  intellect=41  spirit=71  health=17534  mana=20

  Spell Stats:  power=0  hit=9.8%  crit=20.4%  penetration=0  haste=6.4%  mp5=0

  Attack Stats:  power=3965  hit=7.8%  crit=31.3%  expertise=6.8  penetration=0.0  haste=6.4%

  Defense Stats:  armor=974

  Actions:

    bladestorm            Count=  4.0|91.1sec  DPE=21117| 5%  DPET= 3378  DPR= 844.7  Miss=0.0%  TickCount=28  Tick=1916  CritTick=4145|4253|49.5%

    deep_wounds           Count=  0.0| 0.0sec  DPE=    1|16%  DPET=    1  DPR=   0.0  Miss=-1.$%  TickCount=209  Tick=1218

    execute               Count= 18.1|15.5sec  DPE= 6220| 7%  DPET= 3769  DPR= 130.1  Miss=0.2%  Hit=3702  Crit= 8844|12942|49.1%

    heroic_strike         Count= 23.9|10.4sec  DPE= 4511| 7%  DPET=    1  DPR= 375.9  Miss=0.2%  Hit=2379  Crit= 5701| 5806|64.3%

    melee_main_hand       Count= 91.3| 3.3sec  DPE= 2975|17%  DPET= 1279  DPR=   0.0  Miss=0.2%  Hit=1962  Crit= 4245| 4340|49.6%  Glance=1472|23.7%

    mortal_strike         Count= 43.6| 6.7sec  DPE= 4530|12%  DPET= 2746  DPR= 151.0  Miss=0.2%  Hit=2694  Crit= 6452| 6572|49.0%

    overpower             Count= 46.7| 6.3sec  DPE= 5510|16%  DPET= 4792  DPR=1101.9  Miss=0.1%  Hit=2313  Crit= 5535| 5653|99.3%

    rend                  Count= 14.0|21.7sec  DPE= 7901| 7%  DPET= 4786  DPR= 790.1  Miss=0.2%  TickCount=96  Tick=1152

    slam                  Count= 57.4| 4.3sec  DPE= 4026|14%  DPET= 2440  DPR= 268.4  Miss=0.2%  Hit=2390  Crit= 5725| 5842|49.2%


Gains:


    Warrior_T7_MS_Orig:

        anger_management=87.7  (overflow=12.2%)

        avoided_attacks=3.9

        bloodrage=135.8  (overflow=9.5%)

        mh_attack=3591.3  (overflow=23.0%)

        sudden_death=3.7


    Warrior_T7_MS_2ptImpExec:

        avoided_attacks=3.9

        bloodrage=90.9  (overflow=9.1%)

        mh_attack=3731.5  (overflow=20.5%)

        sudden_death=8.8


    Warrior_T7_18_53_00:

        avoided_attacks=47.2

        berserker_rage=125.1  (overflow=4.1%)

        bloodrage=153.7  (overflow=2.5%)

        glyph_of_heroic_strike=507.9

        mh_attack=1276.4  (overflow=4.7%)

        oh_attack=2656.6  (overflow=8.5%)

        unbridled_wrath=141.8  (overflow=10.1%)


Procs:


    Warrior_T7_MS_Orig:

        sudden_death=10.4|28.78sec

        taste_for_blood=48.5|6.18sec


    Warrior_T7_MS_2ptImpExec:

        sudden_death=10.2|29.41sec

        taste_for_blood=48.5|6.18sec


    Warrior_T7_18_53_00:

        bloodsurge=23.8|12.59sec

        darkmoon_greatness=7.0|42.77sec

        mirror_of_truth=5.0|59.77sec


Up-Times:


    Global:

        100.0% : blood_frenzy

        100.0% : mangle

        100.0% : master_poisoner

        100.0% : savage_combat

        100.0% : totem_of_wrath

        100.0% : trauma


    Warrior_T7_MS_Orig:

        54.9% : berserking_mh

        20.8% : heroic_strike

        12.2% : rage_cap


    Warrior_T7_MS_2ptImpExec:

        55.0% : berserking_mh

        20.1% : heroic_strike

        11.1% : rage_cap


    Warrior_T7_18_53_00:

        55.8% : berserking_mh

        29.6% : berserking_oh

        87.7% : flurry

        72.6% : heroic_strike

         5.2% : rage_cap


Waiting:

     0.6% : Warrior_T7_MS_Orig

     0.7% : Warrior_T7_MS_2ptImpExec

    27.1% : Warrior_T7_18_53_00


Baseline Performance:

  TotalEvents   = 5133041

  MaxEventQueue = 36

  SimSeconds    = 299713

  CpuSeconds    = 8

  SpeedUp       = 37464


Share this post


Link to post
Share on other sites

Would it make sense to add criteria similar to Moonkin Druids in regard to the remaining window until the next Taste for Blood refresh (every 6 seconds). (Moonkin have checks such as +=/moonfire,eclipse_left>=10)

For instance, I would say that generally speaking Mortal Strike would logically have priority over Overpower if Taste for Blood had just proc'd and MS is off cooldown, as there is no danger of losing the window of opportunity for Overpower until the proc is due to refresh. Doing Overpower -> MS vs. MS -> Overpower isn't going to change opportunity time for Overpower (provided you use it within the GCD + lag time of the buff refreshing), but it will reduce the opportunity time of MS due to delaying the cooldown refresh.

Share this post


Link to post
Share on other sites
Ok, I just wasn't using the T8 globals file, silly me. Here's what I get with your original and my "better" arms spec, including the imp.execute change I outlined above. Seems that the heroic strike critting has a bigger impact than imp.exec. I wonder if something is wrong or if that is really how it is.

That's how it really is. Imp execute is a pretty bad talent all in all, and Incite destroys it. There's just a lot of lazy/bad/uninformed Arms warriors who think you're not supposed to use HS as Arms for some mystical reason, spread a lot of misinformation about Incite being bad, and you end up with the whole population thinking it's actually true.

Share this post


Link to post
Share on other sites
That's how it really is. Imp execute is a pretty bad talent all in all, and Incite destroys it.

Well, I must disagree. Mainly because you generalise too much assuming everyone has the same/similar very good gear. In fact value of several talents changes together with gear - the better items you have the better some talents become while effectiveness of other lowers.

In case we have here (ImpExecute vs Incite) ImpExecute is better than Incite if you do not have good enough gear to use HS frequently.

Share this post


Link to post
Share on other sites
Well, I must disagree. Mainly because you generalise too much assuming everyone has the same/similar very good gear. In fact value of several talents changes together with gear - the better items you have the better some talents become while effectiveness of other lowers.

In case we have here (ImpExecute vs Incite) ImpExecute is better than Incite if you do not have good enough gear to use HS frequently.

Naxx gear + raid buffs is more than enough to put incite far ahead. I don't think SimCraft is designed to model players in greens/blues and DPS output in heroics.

Share this post


Link to post
Share on other sites

The number of Sudden Death procs seems low in jlassila's results and in my own, unless I misunderstand the mechanics of Sudden Death.

In jlassil's results, there is an average of 309 melee attacks with a roughly 0.02% miss chance, which means that almost all melee attacks were hits. There were an average of 10.2 Sudden Death procs. 10.2 / 309 is roughly 3.3%

Share this post


Link to post
Share on other sites
The number of Sudden Death procs seems low in jlassila's results and in my own, unless I misunderstand the mechanics of Sudden Death.

In jlassil's results, there is an average of 309 melee attacks with a roughly 0.02% miss chance, which means that almost all melee attacks were hits. There were an average of 10.2 Sudden Death procs. 10.2 / 309 is roughly 3.3%

Hmm.. in looking at the code, I'm not 100% familiar with how simulationcraft is set up but I did notice that sudden death is only being triggered at the bottom of the melee hit code. This means it would only proc off melee_main_hand and heroic_strike entries, not every instant attack.

The proc rate seems consistant with those numbers, if that's the case.

That said, I'm not certain it is. Looking at WWS from last night, I'm seeing things like:

0:39'12.563 Jayde melee swing hits Molten Colossus #1 for 3154 Physical. (Critical) #84595

0:39'13.625 Jayde Execute hits Molten Colossus #1 for 6871 Physical. (Critical) #84773

0:39'13.625 Jayde Sudden Death was removed from Jayde. #84774

0:39'13.984 Jayde gains Sudden Death. #84825

0:39'14.391 Molten Colossus #1's Deep Wounds is refreshed. #84874

0:39'14.828 Jayde Overpower hits Molten Colossus #1 for 4903 Physical. (Critical) #84955

0:39'14.969 Jayde melee swing hits Molten Colossus #1 for 1385 Physical. (Glancing) #84977

It seems unlikely that, even with lag, that Sudden Death proc would have been caused by the melee swing 1 second later or 1.4s before.

Also, something like this sequence (Other gains, DoT ticks, etc. clipped out)

0:42'17.984 Jayde melee swing hits Magma Rager #1 for 3184 Physical. (Critical) #99536

0:42'18.797 Jayde Slam hits Magma Rager #1 for 1902 Physical. #99688

0:42'19.172 Jayde gains Sudden Death. #99737

0:42'19.625 Jayde Overpower hits Magma Rager #1 for 4149 Physical. (Critical) #99796

0:42'21.219 Jayde Mortal Strike hits Magma Rager #1 for 5078 Physical. (Critical) #100047

0:42'21.422 Jayde melee swing hits Magma Rager #1 for 1361 Physical. #100070

Seems clear that, at the very least, my Slam or Overpower (Sudden Death applications seem a bit strange in terms of almost always happening 0.3s before or after an attack in the combat log--can't tell which from my current samples) and not any of my swings.

Another example shows:

0:42'54.547 Jayde Heroic Strike hits Magma Rager #2 for 5488 Physical. (Critical) #103819

0:42'54.938 Jayde Mortal Strike hits Magma Rager #2 for 6092 Physical. (Critical) #103869

0:42'55.719 Magma Rager #2's Deep Wounds is refreshed. #103949

0:42'56.531 Jayde Sudden Death was removed from Jayde. #104066

0:42'56.547 Jayde Execute hits Magma Rager #2 for 8148 Physical. (Critical) #104086

0:42'56.969 Jayde gains Sudden Death. #104140

-End of Fight-

0:43'06.953 Jayde Sudden Death was removed from Jayde. #104612

Seems to make it clear that Execute can proc Sudden Death as well.

So, it would seem the proc roll should exist in warrior_attack_t::execute rather than melee_t::execute. (I could be wrong about this though, but quick browsing of the code seems to suggest that.)

Share this post


Link to post
Share on other sites

So, it would seem the proc roll should exist in warrior_attack_t::execute rather than melee_t::execute. (I could be wrong about this though, but quick browsing of the code seems to suggest that.)

Infers possible bug based upon high-level reports.

Confirms bug via careful (and multiple) WWS parses.

Finds bug in source and suggests alternative.

Pinch me! I'm dreaming!

In all seriousness, detailed contributions like this help immeasurably. Ping me if you want to take it that final step and commit the change yourself. Thorough analysis is the fastest way to get me to hand out commit access.....

Share this post


Link to post
Share on other sites

I also noticed in the code that the double-dibbing of Enrages (Wrecking Crew, Death Wish) for Deep Wounds hasn't been fixed.

I tested it earlier today and it doesn't appear to be happening in 3.1. I don't have a parse however. I can work on one if that's needed to make the change.

Share this post


Link to post
Share on other sites
I also noticed in the code that the double-dibbing of Enrages (Wrecking Crew, Death Wish) for Deep Wounds hasn't been fixed.

I tested it earlier today and it doesn't appear to be happening in 3.1. I don't have a parse however. I can work on one if that's needed to make the change.

Oh nice, just tested it, seems to be fixed in 3.1.1. Thanks for the find

Added a run of the sim with the fixed deep wounds and added fancy chart graphics!

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.