I didn't enable mangle/trauma. Nowadays warriors are all Fury and to experiment such an high level gear you will have to be in a high level guild. I doubt you'll find arms warriors there or even 2 cats.

As i said, with Mjolnir i gemmed Agi/Haste on yellow sockets (Deft Ametrite) and Nightmare's Tear on the blue socket of Twin's Pact. That would justify the haste deficiency.

I also enabled Strength of Earth Totem + Enhancing Totems instead of Horn of Winter. That alone is an extra 23 Agi + 23 Strength. As another minor thing, Might instead of Battle Shout will give you +2 Attack Power. Finally, don't forget about Potion of Haste and Bloodlust.

Hope it helped.

what i read over the last few pages is, that there is currently (with 3.2.2) no easy way to say what is best.
first of all, all the different calc tools vary more than the tested combinations. that leads to the question, which is the most accurate?
Another thing is, it seems at the moment, everyone must run the simulator for himself to get "his". before, it was clear, take 2t9 and then take only stuff that has ArPen and so on.

Another thing may be, that with the nerf, blizzard made way for several playstyles? means, if people like a more quiet life, they go for ArPen (no trinket) resulting in more consistent DPS. if you go for softcap+mjolnir and then gem for agi, then your DPS will be less consistent.
For me, the maxed ArPen seems for the current content the better choice, as there is often movement or target switching.
And the other thing: we are now around 10k DPS. these 100 or less DPS difference (1%) on patchwerk-fights make a real difference, not? -> with good movement and so on, you can get far more out of that
Then we have that (seen in the OP) gaussian bell curve DPS spread. i guess the stddev is also higher than 1%.

Conclusion: wether you go for ArPen stacking or for mjolnir+softcap then agi, your playstyle (and skill) matters more than your gear. the longer you stick on a single target (aka the more patchwerk-like the fight gets), mjolnir outweights ArPen stacking
For me, as i dont have mojlnir (sadly) and runnig/switching is a good part in coloseum, ill stay with ArPen stacking.

Changing the softcaped version with agi Stack, with Soul-Devouring Cinch I got this result:
Simulation Nr 1 - Softcap with Agility - with Soul-Devouring Cinch

Slot	Item
Head	Malfurion's Headguard of Triumph (Arcanum of Torment -> 50 AP, 20 Crit Rating)
Neck	Charge of the Demon Lord
Shoulders	Malfurion's Shoulderpads of Triumph (Greater Inscription of the Axe -> 40 AP, 15 Crit Rating)
Chest	Vest of Shifting Shadows (Powerful Stats -> 10 Agility/Stamina/Strengh)
Waist	Soul-Devouring Cinch
Legs	Legguards of the broken Beast (Nerubian Leg Reinforcements -> 75 AP, 22 Crit Rating)
Feet	Treads of the Icewalker (Superior Agility -> 16 Agility)
Wrist	Bracers of Swift Death (Greater Assault -> 50 Attack Power)
Hands	Gloves of the Silver Assassin (Major Agility -> 20 Agility)
Finger1	Dexterous Brightstone Ring
Finger2	Ring of Callous Aggression
Trinket1	Mjolnir Runestone
Trinket2	Death's Choice
Back	Drape of the untamed Predator (Major Agility -> 22 Agility)
MainHand	Archon Glaive (Mungo -> 120 Agility, 2% physical Haste procc)
Idol	Idol of Mutilation

ArP Rating: 730 + Mjolnir Runestone (665 ArP Procc)
Hit Rating: 192
Expertise Rating: 152
Agility: 2674
DPS: 10071

It results in lower DPS.
You kind of trade hit for expertise.
If I got the numbers right, according to nightcrawler hit and expertise are about even in Colusseum BiS. Both come out at a dps value of 7.4
So the overall lower stats of Soul-Devouring Cinch seem to make it the worse choice.

The Vest of Calamitous Fate does not have ArP, while the Armor of Shifting Shadows has 67 ArP rating.
Considering Sylvanna's Cunning I just happened to be unable to find it in Rawr, so I did not include it in the simulation.
Changing the Gear based on this I got this results:

Simulation Nr 1 - Softcap with Agility - with changed Items

Slot	Item
Head	Malfurion's Headguard of Triumph (Arcanum of Torment -> 50 AP, 20 Crit Rating)
Neck	Charge of the Demon Lord
Shoulders	Malfurion's Shoulderpads of Triumph (Greater Inscription of the Axe -> 40 AP, 15 Crit Rating)
Chest	Curiass of Calamitous Fate (Powerful Stats -> 10 Agility/Stamina/Strengh)
Waist	Belt of the Merciless Killer
Legs	Legguards of the broken Beast (Nerubian Leg Reinforcements -> 75 AP, 22 Crit Rating)
Feet	Treads of the Icewalker (Superior Agility -> 16 Agility)
Wrist	Bracers of Dark Determination (Greater Assault -> 50 Attack Power)
Hands	Gloves of the Silver Assassin (Major Agility -> 20 Agility)
Finger1	Dexterous Brightstone Ring
Finger2	Ring of Callous Aggression
Trinket1	Mjolnir Runestone
Trinket2	Death's Choice
Back	Drape of the untamed Predator (Major Agility -> 22 Agility)
MainHand	Archon Glaive (Mungo -> 120 Agility, 2% physical Haste procc)
Idol	Idol of Mutilation

ArP Rating: 661 + Mjolnir Runestone (665 ArP Procc)
Hit Rating: 289
Expertise Rating: 96
Agility: 2705
DPS: 10237


This part is nonsense, due to my misunderstanding of Cat-Swipe. Lupine Longstaff/Twin's Pact is always better then Archon Glaive.
Considering the weapons:
In this simulation Lupine Longstaff/Twin's Pact comes out 3520,22 Points in the Item Comparsions, while Archon Glaive come out at 3508, 35.
Changing the Weopons and gemming Lupine Longstaff/Twin's Pact with ArP rating the DPS rises to 10277.
Lupine Longstaff/Twin's Pact is the better weapon in terms of single target DPS.
Thoug I think it sould be pointed out here that Lupine Longstaff/Twin's Pact lacks AE power.
Since Cat-Swipe scales with weapon damage and Lupine Longstaff/Twin's Pact does only have 540 - 811 Damage while the Archon Glaive has 788 - 1183 Damage.
In encounters where AE matters (in example Freya) Archon Glaive would still be better.
Of course the best way would be to have both of them.
This part is nonsense, due to my misunderstanding of Cat-Swipe. Lupine Longstaff/Twin's Pact is always better then Archon Glaive.


@Odas
Like I said to ramenchef, I can't get Sylvanna's Cunning / Veressa's Dexterity into my Rawr.
Also how do I tell my Rawr to use both proccs form the 245 as well as the 258 version of Death's Choice?
I tried recreating some of your simulations and got different results, but righ now I am assuming this is due to the fact that I do not use Sylvanna's Cunning / Veressa's Dexterity.
I assume it would be pretty much pointless posting all the simulations that are off, until I can get Sylvanna's Cunning / Veressa's Dexterity to work.
Again I used all available raid buffs with the expectation of the draenei hit aura, gains from professions, double pot trick, 80 Attack Power gain from Mixology and the resistance buffs.

I still wanted to post the first one I tried:


Softcap + Mjolnir + 4xT9

Slot	Item
Head	Malfurion's Headguard of Triumph (Arcanum of Torment -> 50 AP, 20 Crit Rating)
Neck	Charge of the Demon Lord
Shoulders	Malfurion's Shoulderpads of Triumph (Greater Inscription of the Axe -> 40 AP, 15 Crit Rating)
Chest	Malfurion's Raiment's of Triumph (Powerful Stats -> 10 Agility/Stamina/Strengh)
Waist	Belt of the Merciless Killer
Legs	Malfurion's Legguards of Triumph (Nerubian Leg Reinforcements -> 75 AP, 22 Crit Rating)
Feet	Treads of the Icewalker (Superior Agility -> 16 Agility)
Wrist	Bracers of Dark Determination (Greater Assault -> 50 Attack Power)
Hands	Gloves of the Silver Assassin (Major Agility -> 20 Agility)
Finger1	Dexterous Brightstone Ring
Finger2	Ring of Callous Aggression
Trinket1	Mjolnir Runestone
Trinket2	Death's Choice
Back	Drape of the untamed Predator (Major Agility -> 22 Agility)
MainHand	Lupine Longstaff (Berserking -> 400 Attack Power procc)
Idol	Idol of Mutilation

ArP Rating: 701 + Mjolnir Runestone (665 ArP Procc)
Hit Rating: 279
Expertise Rating: 86
Agility: 2581
Attack Power: 13047
Haste: 578
Crit Rating: 756
Strengh: 371
DPS: 10320

My simulation comes out with 272,85 less DPS then your simulation.
Again I am not using any gains from professions here.
using 3 fractured dragon's eyes I would still land at 10399 DPS, what would still be 193,85 DPS lower than on your version.
Are those numbers the result of the missing Sylvanna's Cunning / Veressa's Dexterity?

Also notice that I am assuming your icewalker enchant on the Icewalker Treads is a typo, since with 279 hit rating (I get the same hit on my simulation) you are already 16 Hit over the cap, making Superior Agility -> 16 Agility clearly the better enchant.

Is there any chance someone could run the FbN simulator using the new ArP values so we can get a comparison to these numbers from Rawr?

I've never trust Rawr too much since it is just a calculator, not a simulator, and, correct me if I'm wrong, it still doesn't properly calculate our CP generation because it just averages crits/energy.

Right click on any cloak, edit and reform it so it has the exact same stats as Sylvanna's Cunning.

I believe the reason Ramenchef said Vest of Calamitous Fate, regardless of softcap or hardcap is that at least when i ran Rawr, in every case scenario, the stats on a no-ArP ilvl258 item outweighted the stats on a ilvl245 ArP item. And since nowadays you can't get over 1142 ArP or so, the ilvl245 items don't become good enough to outweight the stats on the ilvl258 ones. Please correct me on this one if i'm wrong.

Select Darkmoon Card: Greatness, swap the 90 agility on it for 256 AP, change the buff to 450 Agi and put the proc at 35% if it isn't already.

In order to achieve the highest results you'll have to choose BS and JC. With those 2 you can't select the 80 AP from Mixology as that's another profession.

Thank you for the help and clarifications.
I will upgrade all my simulations taking Sylvana's Cunning/Vereesa's Dexterity into the sims.

Considering the professional gains I realise that JC will come out top.
I still am hesistant to use the in my simulations as I want to keep them more general and not assume certain professions.
Of course I do realise that going for maximum DPS they will be necessary.

You can build a gear with hit and expertise capped with [Soul-Devouring Cinch].

Slot	Item
Head	[Runetotem's Headguard of Triumph]
Neck	[Charge of the Eredar]
Shoulders	[Runetotem's Shoulderpads of Triumph]
Chest	[Vest of Calamitous Fate]
Waist	[Soul-Devouring Cinch]
Legs	[Legwraps of the Broken Beast]
Feet	[Treads of the Icewalker]
Wrist	[Bracers of Swift Death]
Hands	[Gloves of the Silver Assassin]
Finger1	[Planestalker Band]
Finger2	[Ring of Callous Aggression]
Trinket1	[Mjolnir Runestone]
Trinket2	[Death's Choice]
Back	Sylvannas Cunning
MainHand	[Twin's Pact]
Idol	[Idol of Mutilation]

Icewalker on Feet / Berserking on Weapon
4 [Deadly Ametrine] on some slots for socket bonus (Head, Shoulders, Finger1 and Feet)
10 [Fractured Cardinal Ruby]
5 [Delicate Cardinal Ruby]
[Nightmare Tear] for Meta Socket activation
ArP JC prismatic gems

ArP Rating: 734 + Mjolnir Runestone (665 ArP Procc)
Hit Rating: 267
Expertise Rating: 142
Agility: 2701
DPS: 10305

Every buff, except Bloodlust and Trauma/Mangle.
ArP food.

[Twin's Pact] > [Hellion Glaive] with any gear.

If I'm not mistaken, when swipe refers to "weapon damage" it does not mean the actual item, but rather our cat form "weapon" that we actually swing with, which is still present and quite separate.

@ Furial, your hardcap ArP set differs very slightly from your softcap, 900 ArP is not hardcap, nor is it even close. However, after some extensive testing with gear sets and as using as many different combinations I could think of, I found that pretty much every set of gear I made, whether hardcap/softcap or 2pc/4pc, differed very little in calculated DPS, with a range of +-100 DPS for each set. Assuming the calculations are close to accurate, would it be safe to say that whichever benefits your playstyle and preference more is equally viable and useful as any other?

Also, as a side note, your comment about arms warriors in high end guilds is vastly incorrect. They can do nearly the same amount of DPS as fury warriors, only falling behind on AoE damage. However, if a guild runs with a feral, the buff that they can provide easily overcompensates for the slight DPS loss they have as fury. Arms warriors are incredibly viable if they are extremely skilled. We run with both a fury and arms warriors, who compete in DPS constantly. Our arms warrior is in the top 5 DPS of every fight and hardmode that we have done for his class/spec on WMO.

Yes, its from "weapon dmg" in cat form. 250% of swing in cat form, not 250% of the Weapon swing.

@ Talanik: You misunderstood me. I'm aware the hardcap is at 1400 ArP. What i meant with "Hardcap" was more an "attempt to reach hardcap"which was basically no Mjolnir and full ArP gemming. Inside this "attempt to reach hardcap" idea 2 sims were made; with ilvl245 ArP gear and with no ArP ilvl258 items.

mod - remove this post please

I've read all the pgs since 3.2.2 and the question has been asked a few times now, but no response.
What is the new Stats value for ArP? The last time it was updated in Aug was 2.74. Since 3.2.2, ArP was nerfed 12%, I assume it not simply decreasing the ArP stats value by 12%, as Agi stats value will come up too?

I've been thinking about the effectiveness of being Soft-capped + mjolnir and have come to the following conclusions:

First, I assume that a reasonable approximation of the damage done with mjolnir would be

TotalDPS = NoProcDPS*(100-Uptime)/100 + ProcDPS*(Uptime)/100

This leads to

TotalDPS = NoProcDPS + (ProcDPS-NoProcDPS)*Uptime/100

If you take the derivative of TotalDPS with respect to MjolnirUptime you get

d(TotalDPS)/d(Uptime) = (ProcDPS-NoProcDPS)/100

The two interesting paths to take from here are either find what the change in DPS is from a change in the internal cooldown(CD) or a change in the duration.

Uptime = 100*(Duration/CD)

d(Uptime)/d(CD) = -100Duration/CD^2

d(Uptime)/d(Duration) = 100/CD

You can then multiply these together to get what the change in DPS would be for different changes in either CD or Duration.

d(TotalDPS)/d(Uptime) * d(Uptime)/d(CD) = d(TotalDPS)/d(CD) = -(ProcDPS-NoProcDPS)*Duration/(CD^2)

d(TotalDPS)/d(Uptime) * d(Uptime)/d(Duration) = d(TotalDPS)/d(Duration) = (ProcDPS-NoProcDPS)/(CD)

I used the profile I listed on the previous page under Softcap + Mjolnir + 4t9 including heroic presence, race changed from Tauren to Night Elf, and trauma (it turns out the difference isnt significant using the profile posted on the previous page exactly). This gave me for values:

TotalDPS: 10722
NoProcDPS: 10356
ProcDPS: 12017
Uptime: 22.03%
Duration: 10s
CD: 45.4s

Using those values:

d(TotalDPS)/d(CD) = -8.06 dps

d(TotalDP)/d(Duration) = 36.59 dps

Considering the differences we have been dealing with between Soft-capping and aiming for the hard cap are on the order of 50dps it seems like an unlucky Mjolnir proc could easily swing you in the other direction. Note that while the CD could effectively be lower than 45s (while you're waiting for the CD you're doing a phase transition or something where you cant attack anyway) the duration wont ever be longer than 10s. Even under the situation when you are low on energy, or are reapplying your bleeds and roar so you don't get to shred much the value of Mjolnir would decrease significantly.

Please let me know if you find any issues with my assumptions or math.

If we're going to factor in wasting a proc for being away from a mob, or having to use other abilities than shred during one, then you have to apply that to death's choice (or every other proc trinket) as well since it has the same problem.

True; however, both of the sets use Death's Verdict so that would effectively cancel out, and other trinkets have smaller variations based on cooldown and duration changes.

Death's Verdict (ilevel 258 Softcap+Mjolnir):

d(TotalDPS)/d(CD) = -7.02 dps

d(TotalDPS)/d(Duration) = 21.07 dps

Death's Verdict (ilevel 258 Hardcap):

d(TotalDPS)/d(CD) = -7.06 dps

d(TotalDPS)/d(Duration) = 21.17 dps

Death's Verdict (ilevel 245 Hardcap):

d(TotalDPS)/d(CD) = -6.24 dps

d(TotalDPS)/d(Duration) = 18.71 dps

Not only does Death's Verdict suffer less from unlucky procs due to not being as powerful of a proc, but also because it has a longer duration (losing 1 second on a 10 second proc is much worse than losing 1 second on a 15 second proc).

But what about lucky procs? While Mjolnir suffers more from unlucky procs, it also benefits more from lucky ones.

Let's say your 22% uptime is correct in the infinite attack sequence scenario. If that's true, then you're more likely, at any random point, to be missing trinket downtime instead of uptime. This remains true until trinket uptime reaches ~50%.

However, trinket procs also have the added benefit of being more likely to occur right after a period of downtime; in other words, if a trinket comes off of cooldown while you aren't attacking, then it's going to pop when you start attacking again, guaranteeing uptime against the boss.

Finally, any real-world scenario has higher trinket uptimes than 22% anyway-- I assume you just took 10/45 and assumed that was reasonable. Like the above scenario, however, you have a beginning-of-fight bias-- your first proc doesn't happen at a random point during the first 45 seconds of the fight, it happens within the first 5-10 seconds. Realistically, you'll have uptimes around 25-30%.

Your numbers are probably off the wrong way-- the larger the proc, the more you want it in a fight that gets interrupted, because you are more likely to get a higher uptime in those cases. And procs tend to have higher uptimes during your DPS period than you may think.

It kinda averages out though, Allev. While it's true that it's more likely for a fight to end before the cooldown on your trinkets comes back up, it's not going to change much when you're dealing with 5+ icd cycles. Your Best Case Scenario has procs immediately after the icd and a fight that lasts exactly long enough for your last proc to finish and no longer, the average boss is either a 3-4 min or 6-7 min fight, so we're looking for a 190 second fight with 5 procs and a 370 second fight with 9 procs. This gives you 26.31% and 24.32% uptimes respectively.

Now, effective uptimes based on DPS Time can go up more with a phase shift immediately after the buff wears off, but the uptimes can also go down a lot more if the change happens immediately just before or just after you get the proc. Big Bang, for instance, is every 90 seconds, if you aren't paying attention, it's entirely possible to proc just before running into a black hole, the best case here is that have to wait til the 100th second at the soonest before you can proc again. You're still most likely to still get 8 procs if you run the full 360 seconds, but, as an example, The last Algalon kill we logged was a 349 second fight and not a single person using a proc trinket only got 7 procs except 1 who got 6 and 1 who got 8 on his DMG. That drops them down to 20% uptime, I can't tell effective uptime vs DPS time, but taking 27 seconds off puts it at 21.74% uptime. I know it's completely anecdotal, but with all but 1 person failing to meet the 8 proc/6 min assumption.

The most amusing part is that even missing 1 proc, it still ends up better uptime than any "use" trinket.

Let's try ballparking DPS time by looking at Neemu's autoattacks. By his current armory, Neemu's haste is at 489, and with appropriate buffs the average swing speed he has should be .6755 (assuming Toskk's is doing the math right, which is a pretty safe bet) and he got 439 swings, 439 x .6755 = 296.5 seconds of attack. But that calc uses heroism averaged over a continuous 360 second fight, so Neemu had even less DPS time than that. So I'd put the range of downtime around 50+ seconds instead of 27.

Which puts uptime over 23% with 7 procs, assuming the procs never got interrupted, which is a safe bet for pretty much anyone but the player with 8 procs.

Also, you ignored the time to proc after it's up from cooldown-- I meant to correct this in the earlier post but did not. While it can proc immediately, the average case for ferals is in the realm of 3-5 seconds (depending on the trinket and whether it's an on-hit or on-crit). And ferals have an advantage over many classes with its fast swing timer-- if you assume on average for classes it's 5 seconds between procs, then 350s is the exact wrong amount of time for a fight. You only expect 7 procs with roughly 50s between them. With the added bonus that you can't proc between 90s and 100s (more or less), you couldn't get ahead with lucky procs on Algalon unless you got 3 consecutive procs within a total of (however long you can DPS after the 90s mark-- I haven't done the Algalon fight, so I don't know). It's possible, but I imagine it's about as rare on that fight as your guild demonstrated.

Impressive job of finding the exception that proves the rule, though!

@Curan
I just don't like to see Dodges or Misses on my screen. That's why I aways look after Hit and Expertise cap. I think if u trade Soul-Devouring Cinch for Belt of the Merciless Killer on my set you get 6-7 more dps, but I prefer to not have Dodges or Misses than 6-7 dps.

I've not seen any tests with Banner of Victory+Death's Choice yet, how come?

Even after the ArP nerf I'm still able to hit 85% passive ArP, which is where the breakpoint is supposed to be. Again tho, all the tests showing 10500 dps and above are totally utopic. I'm very close to having the BiS gear for hardcapping ArP and I barely do 9400 dps on Jaraxxus (not changing target). Somehow I dont see 1 trinket, a cloak and 1 245 > 258 upgrade giving me 1.1k extra DPS. So my question is.. how much of that is down to logical flaws in the program and how much is the players fault? I've never used Rawr, FBN or anything else; I've always used my own logic. Thats solely because I dont trust Rawr, FBN or whatnot because the numbers generated by those programs are ridiculous in all respects.

What programs like Rawr and FBN calculates is Maximum theoretical dps. In-game will always be less than that due to a number of reasons(lag, movement, buffs not being present 100% of the time and what not), and there might be a few flaws in regards to how trinkets and other procs are calculated, since we don't have full access to game-code. That does not invalidate the stat values or BiS gear list though, and it will most certainly give a better picture than common-sense logic.