Kahorie's DK simulator 2.0 Beta - Including Cataclysm. - Page 19

dr_AllCOM3 · 09/17/09, 8:54 AM

All DK specs have so few free talent points, that doing them manually is much faster. Additionally there are far too many possible combinations to test.

Afabar · 09/17/09, 9:02 AM

I have made some improvements to the MThreading support. You will now have a progress bar for each simulation. Note that a full EP stat calculation is multiple simulations so don't be afraid of the number of progress bar that will pop.
I have also set thread priority to low to avoid to drown your computer.

Kahorie's DK Simulator 1.1.1.1 MT.zip

Doc, If you have questions regarding the code reorganization, fell free to ask.

GravityDK · 09/17/09, 9:24 AM

Progress bar works, nice.

If I load the app, start a DPS sim, then close the app during the simulation via close-box in top right, the app gives crash.

Description:
Stopped working

Problem signature:
Problem Event Name: CLR20r3
Problem Signature 01: kahorie's dk simulator.exe
Problem Signature 02: 1.1.1.1
Problem Signature 03: 4ab23068
Problem Signature 04: Microsoft.VisualBasic
Problem Signature 05: 8.0.0.0
Problem Signature 06: 49cc7d22
Problem Signature 07: 359
Problem Signature 08: 52
Problem Signature 09: System.InvalidCastException
OS Version: 6.0.6002.2.2.0.768.3
Locale ID: 2057

Also, sadly and frustrating, if I use my old xml files (priority/template/character) I still get a crash. If this is not fixable, that's OK, I'll have to recreate but it'd be a nice time-saver if there was a solution.

Description:
Stopped working

Problem signature:
Problem Event Name: CLR20r3
Problem Signature 01: kahorie's dk simulator.exe
Problem Signature 02: 1.1.1.1
Problem Signature 03: 4ab23068
Problem Signature 04: Kahorie's DK Simulator
Problem Signature 05: 1.1.1.1
Problem Signature 06: 4ab23068
Problem Signature 07: 62
Problem Signature 08: 4c2
Problem Signature 09: System.NullReferenceException
OS Version: 6.0.6002.2.2.0.768.3
Locale ID: 2057

Consider · 09/17/09, 12:20 PM

Is it possible to add the EP calculations a "X DPS per 1.00 AP" (Or, vice versa - 1 DPS per X.YZ AP")? It should be fairly easy, if I'm not mistaken, since the sim already does a dry run and an AP run.

The current stat weights are relative to AP, which is great for making gearing/gemming/consumable decisions. However, it doesn't actually tell you how much of a dps upgrade an item is or how well a spec scales relative to other specs. The latter is of interest to me, although I'm sure there are those who would be interested for the first reason as well.

GravityDK · 09/17/09, 1:43 PM

The latter interests me a LOT, also. Tank spec decision making is full of trade-offs because we invest 18 talents points in tanking, not dps, talents. So spec v spec analysis tools, yes please.

Mericet · 09/17/09, 4:33 PM

I'm having a problem with the "many fight" button on EP simulations. It seems to run alright, but for some reason it screws up the haste EP calculation. When I run the sim without the many fights button, I get a haste value of 1.8 (by the way, is there a way to make version 1.1.1 report 3 figures like the older ones? I assume it rounds, but I'd still rather have 1.84 than 1.8), but when I turn that setting on it gives me haste values of .3. I didn't change anything else between simulations, and I tried with 1.0.9 as well and found the same issue. Let me know if there's any other information that could help and I'll edit it into this post. The simulations were done as 51-0-20 blood (similar to the preset but not exactly the same) with the blood priority setting.

Raef1 · 09/19/09, 3:30 AM

Is the 4 piece T9 template option working correctly?
I have it selected (by replacing the 0 with a 1) for my simulation but the results are showing my diesease crit% as 0.

Am I doing something wrong?

dr_AllCOM3 · 09/19/09, 10:48 AM

I've added aoe damage in the SVN version. It's not totally polished (Stuff that only affects the main target e.g. and diseases are treated as always up on every target), but it will be interesting for talent or item decisions.

Afabar · 09/19/09, 5:39 PM

Gravity, I thinks I have found what is wrong with your templates files, they are missing the new trinket section. It should be corrected in the next release.

Consider, it should not be too complex to add. I will look at it.

Raef1, the T9 P4 bonus is implemented. It's just the report that don't show the crit percentage.

In the next release, it will be include a new talent editor to get rid of MMO-champion changes.
Edit: New release:
1.1.2 Release notes
* Expertise value after dodge cap calculation on Frost presence.
* A "none" option in Runeforge and Sigil options
* Rune Strike benefit from ToT
* Multithreading support.
* Internal Talent calculator.
* Multitarget implementation

Direct link: Kahorie's DK Simulator 1.1.2

GravityDK · 09/20/09, 1:12 PM

Regarding 1.1.2:
a) my existing xml files work in this version (THANK YOU),
b) multithreaded progress bar works
c) load appears to be across all CPUs, albeit unevenly [screenshot]
d) 'none' sigil and runeforge tested on & off, they work too, (thanks again).

Catch: still spams RS in Frost Presence, distorting the DPS results badly.

I appreciate the start you've made on multi-target module, that will be useful. I guess by writing a multi-target priority file I could make it useful already. Suggestion: include number of targets in the Report page, to eliminate uncertainty.

Thank you two so much for all this work. Brilliant contribution.

Bug?: I wonder if there's a bug in this version which inflates DPS results. On the same settings I've used before, I now get 4500 dps rather than about 3000.

UPDATE: it was buffs, not a bug. The new version put all the raid buffs back on. 50% increase from having every raid buff up, wow, that's massive.

In the 1.1.2 version:

Ability Total % Landed Hit% Crit% Miss% Average
MainHand 412967192 25.1 132563 68.8 31.2 0 3115.3
ScourgeStrike 338428011 20.5 63686 63 37 0 5314
DeathCoil 270745399 16.4 51956 71.9 25.3 2.8 5211.1
FrostFever 154210742 9.4 108836 100 0 0 1416.9
BloodPlague 141561795 8.6 107142 100 0 0 1321.3
Necrosis 82556472 5 132563 100 0 0 622.8
BloodStrike 80621325 4.9 31596 67.9 32.1 0 2551.6
UnholyBlight 75777847 4.6 51956 100 0 0 1458.5
PlagueStrike 45308044 2.8 15941 63 37 0 2842.2
IcyTouch 45104079 2.7 15549 71.9 25.4 2.7 2900.8
Horn 0 0 15354 100 0 0 0
DPS 4576
Total Damage 1647.28m in 100h

Whereas this is result from test in 1.1.1, same build/spec/etc BUT WITH BUFFS:

Ability Total % Landed Hit% Crit% Miss% Average
Scourge Strike 272349774 24.6 63488 72 28 0 4289.8
Main Hand 231642789 21 105883 78.3 21.7 0 2187.7
Death Coil 173652114 15.7 49917 81.3 11.9 6.8 3478.8
Frost Fever 112750094 10.2 111066 100 0 0 1015.2
Blood Plague 103507376 9.4 111002 100 0 0 932.5
Blood Strike 56112356 5.1 31740 78.1 21.9 0 1767.9
UB 48634234 4.4 49917 100 0 0 974.3
Necrosis 46307783 4.2 105883 100 0 0 437.3
Icy Touch 30568475 2.8 15867 81.1 11.6 7.3 1926.5
Plague Strike 30141269 2.7 15880 72.1 27.9 0 1898.1
Horn 0 0 15857 0 0 0 0
DPS 3071
Total Damage 1105.67m in 100h

dr_AllCOM3 · 09/20/09, 1:19 PM

Originally Posted by GravityDK

c) load appears to be across all CPUs, albeit unevenly [screenshot]

Only EP is multithreaded.

GravityDK · 09/20/09, 1:28 PM

Originally Posted by dr_AllCOM3

Only EP is multithreaded.

Oh wow, yeah I can see that now. Very nice. Screenshot FYI:

Mind · 09/21/09, 10:04 AM

After doing some tests I come to a conclusion that, for example, +50 ap bracers doesn't add "green" attack power to chart, instead it increases normal attack power. Does it influence the simulator since there are only str and green ap values on template?

Afabar · 09/21/09, 10:27 AM

I wasn't aware of that and so is the simulator. I have to think of a walk around.

Nihillus · 09/21/09, 5:01 PM

PROBLEM RESOLVED EDIT: This post was out of ignorance and my error, feel free to delete. I didnt understand how the rotation file worked at the time. Thanks Gravity.

I'd like to test an unholy DW build like 0/13/58 using a rotation like:

IT > PS > SS > Dump > BS > BS (75 RP)
IT > PS > SS > Dump > GF > IT (85 RP)
IT > PS > SS > Dump > BS > BS (75 RP)
IT > PS > SS > Dump > IT > IT (95 RP)

Basically the theory of this build/rotation is kind of like the DW spam build, but using SS to keep up virulence and thus modifying the build to make the best of SS.

It seems to me the current talk in the 2h UH thread are changing more towards this end (using SS over OB, dropping epidemic, some talk of UP being better on single target), and I would like to run it's numbers to compare. However there is a problem with testing this build.

The rotation only really uses SS to maintain the Virulence buff (freeing a GCD is a nice bonus for bursts or outside RP gains). As such a priority would be something like:

Diseases > MaxRP > Virulence > Ghoul Frenzy > IT > PS > BS > DC > Horn

However, there is no priority to SS for Virulence in the simulator atm. This is the main point of this post.

Additionally, the ICD of Virulence is 10sec iirc, reducing it's uptime in a IUP rotation (9sec) and complicating an optimized priority system for players.

Another issue is that I don't think the simulator refreshes diseases once per cycle unless the application skills are called in the priority.

Ideally, I think a rotation is more user friendly and practical to use than a priority for this build, but I realize that the program is priority based. As such I may just have to try to test the numbers manually.

edit: pardon the length of the non-underlined, but i was going to post this in the 2h uh thread before deciding the heart of the issue was more relevant to this thread. i edited out some parts that were non-essential to this thread.

edit2: stuck in maxrp and dc to the priority for completeness

GravityDK · 09/22/09, 5:33 AM

I'm not clear why the rotation file would not work in your test scenario? You can tell it to use a rotation rather than priority.

GravityDK · 09/22/09, 7:43 AM

I just did EP calculations of the same spec/priority/gear etc but changed from blood to frost presence. Because of the way rune strike now comes into it, spamming, it's overly weighting haste and down-rating crit. Just FYI, one of the inaccuracies I'm fairly sure are a consequence of the current RS simulation.

Results:

ron9 · 09/22/09, 5:23 PM

I'm using 12-0-59 spec with Priority of:

<?xml version="1.0"?>
<Priority>
<BloodPlague></BloodPlague>
<FrostFever></FrostFever>
<DeathandDecay></DeathandDecay> (underlining for emphasis)
<ScourgeStrike></ScourgeStrike>
<BloodStrike></BloodStrike>
<DeathCoil></DeathCoil>
</Priority>

and EP values are consistently showing that Haste makes me lose DPS by valuing it at a negative value when I have enemies set to 2+. I have ran the simulation at 100+ hours.

I suspect that when running a Single Target DPS test with 12-0-59 and the default Unholy Priority that Haste is being severely under-valued. In the DPS simulator version, updated with all the current 3.2.2 changes, a few versions ago I was getting Haste EP values that were nearly as good as the Hit EP values.

Afabar · 09/22/09, 6:22 PM

I suppose it's a spec without Epidemic? If it's true, then it don't surprise me. The sim do not like priority with 15 second disease. I have try with 2 point in epidemic:

EP:50	AttackPower	1
EP:50	Strength	2,93
EP:50	Agility	0,96
EP:50	CritRating	2,11
EP:50	HasteRating	2,25
EP:50	ArmorPenetrationRating	1,01
EP:50	ExpertiseRating	1,15
EP:50	BeforeMeleeHitCap<8%	3,86
EP:50	SpellHitRating	3,03

ron9 · 09/22/09, 6:52 PM

With the default 12-0-59 with default Unholy priority and 100hour simulation I got:

EP:50 | AttackPower | 1
EP:50 | Strength | 3.05
EP:50 | Agility | 1.3
EP:50 | CritRating | 1.9
EP:50 | HasteRating | 0.22
EP:50 | ArmorPenetrationRating | 1.05
EP:50 | ExpertiseRating | 1.49
EP:50 | BeforeMeleeHitCap<8% | 2.76
EP:50 | SpellHitRating | 0.98
EP:10 | WeaponDPS | 6.03

I am using 4set T9.

Here is the problem, I am using these Stats for Character profile:

<?xml version="1.0"?>
<character>
<stat>
<Strength>2050</Strength>
<Agility>41</Agility>
<Intel>58</Intel>
<Armor>14996</Armor>
<AttackPower>503</AttackPower>
<HitRating>370</HitRating>
<CritRating>807</CritRating>
<HasteRating>262</HasteRating>
<ArmorPenetrationRating>300</ArmorPenetrationRating>
<ExpertiseRating>203</ExpertiseRating>
</stat>
<weapon>
<count>1</count>
<mainhand>
<dps>255</dps>
<speed>3.5</speed>
</mainhand>
<offhand>
<dps>0</dps>
<speed>0</speed>
</offhand>
</weapon>
<EP><base>50</base></EP>
<Set>
<T72PDPS>0</T72PDPS>
<T74PDPS>0</T74PDPS>
<T82PDPS>0</T82PDPS>
<T84PDPS>0</T84PDPS>
<T92PDPS>1</T92PDPS>
<T94PDPS>1</T94PDPS>
<T72PTNK>0</T72PTNK>
<T74PTNK>0</T74PTNK>
<T82PTNK>0</T82PTNK>
<T84PTNK>0</T84PTNK>
</Set>
<ChaoticSkyflareDiamond>1</ChaoticSkyflareDiamond>
<trinket>
<MjolnirRunestone>0</MjolnirRunestone>
<GrimToll>0</GrimToll>
<BitterAnguish>0</BitterAnguish>
<Mirror>0</Mirror>
<Greatness>0</Greatness>
<DCDeath>0</DCDeath>
<Victory>0</Victory>
<Necromantic>0</Necromantic>
<Bandit>0</Bandit>
<Pyrite>0</Pyrite>
<DarkMatter>0</DarkMatter>
<OldGod>0</OldGod>
<Comet>0</Comet>
<DeathChoice>0</DeathChoice>
<DeathChoiceHeroic>0</DeathChoiceHeroic>
</trinket>
</character>

Is there a reason why this would make Haste so de-valued?

Fearlezz · 09/23/09, 8:27 AM

dr_AllCOM3 · 09/23/09, 8:35 AM

Originally Posted by Fearlezz

Calculating trinket EP always gives me this result:

Your answer is here.

Fearlezz · 09/23/09, 8:51 AM

Originally Posted by dr_AllCOM3

Your answer is here.

Now that you linked it I remember reading it /facepalm.

On a side note, did anyone else have similar results with Blood ArP calculations (post nerf)? Posted mine here (link). I used /dksim (addon) to export character stats to the simulator and if the results are correct, gemming ArP with my set of gear is still suboptimal even at 800ish ArP rating on gear.

Raika · 09/24/09, 2:21 PM

I'm having trouble getting the same number of strikes out of two rotations. Both of these should have the same number of blood strikes, but I'm consistently getting fewer out of the first. I'm assuming it's something I'm doing wrong.

Using:
0300000000000000000000000000320053000000000000000000000002300303350032152000150003133 151&glyph=270817&

<?xml version="1.0"?>
<Rotation>
	<Rotation>
	<PlagueStrike retry='1'></PlagueStrike>
	<IcyTouch retry='1'></IcyTouch>
	<ScourgeStrike retry='1'></ScourgeStrike>
	<BloodStrike retry='1'></BloodStrike>
	<BloodBoil retry='1'></BloodBoil>
	<DeathCoil retry='0'></DeathCoil>
	<DeathCoil retry='0'></DeathCoil>
	<GhoulFrenzy retry='1'></GhoulFrenzy>
	<IcyTouch retry='1'></IcyTouch>
	<ScourgeStrike retry='1'></ScourgeStrike>
	<BloodStrike retry='1'></BloodStrike>
	<BloodBoil retry='1'></BloodBoil>
	<DeathCoil retry='0'></DeathCoil>
	<PlagueStrike retry='1'></PlagueStrike>
	<IcyTouch retry='1'></IcyTouch>
	<ScourgeStrike retry='1'></ScourgeStrike>
	<BloodStrike retry='1'></BloodStrike>
	<BloodBoil retry='1'></BloodBoil>
	<DeathCoil retry='0'></DeathCoil>
	<DeathCoil retry='0'></DeathCoil>
		</Rotation>
	<Runes>
			<Blood></Blood>
			<Blood></Blood>
			<Frost></Frost>
			<Frost></Frost>
			<Unholy></Unholy>
			<Unholy></Unholy>
		</Runes>
</Rotation>

vs

0300000000000000000000000000320053000000000000000000000002300303350032152000150003133 151&glyph=270817&

<?xml version="1.0"?>
<Rotation>
	<Rotation>
	<PlagueStrike retry='1'></PlagueStrike>
	<IcyTouch retry='1'></IcyTouch>
	<ScourgeStrike retry='1'></ScourgeStrike>
	<BloodStrike retry='1'></BloodStrike>
	<BloodBoil retry='1'></BloodBoil>
	<DeathCoil retry='0'></DeathCoil>
	<DeathCoil retry='0'></DeathCoil>

		</Rotation>
	<Runes>
			<Blood></Blood>
			<Blood></Blood>
			<Frost></Frost>
			<Frost></Frost>
			<Unholy></Unholy>
			<Unholy></Unholy>
		</Runes>
</Rotation>

Blood Presence, Sigil of Virulence, FC/FC, 100 hour sim

The first gives 32987 blood strikes, the second 33183. All the strike amounts seem to be less with the first build/rotation.

EDIT: Tried to make it a little cleaner with code tags.

Afabar · 09/24/09, 5:47 PM

It comes from <GhoulFrenzy retry='1'></GhoulFrenzy>
The sim already use it automatically in the first rotation. So it's unavailable a few second later. As a walkaround set to retry='0', it should to the trick after, the rotation will tune itself.
However I will correct this in the next version, but before delivering it I want to implement the new Bone Shield cooldown but I have to found a solution for it to be efficient. Right now, depending of rotations or priorities, it can be a dps loss so I need to work on it.