Your question can be answered in this chart from the previous page, i believe.

The 2piece is absolutely amazing, much easier to keep up finishers with the additional shreds. (3 back-to-back ooc procs in 5seconds while having storm power at hodir = win)

You should better use two piece of T8 gear, ooc proc is an amazing burst dps.
Head and Gloves / Shoulders is the best.

Idol of the Ravenous Beast or Idol of Worship if not stacking ArPen?

worship

Despite having ~545 arp and both Mjolnir and GT, I'm pulling more damage with Idol of Worship than with the Shred idol. I've got no idea why.

Do you have to mangle?

I'm running with another feral druid, and whoever's got the energy to mangle when the debuff runs out gets to do it. I'd guess there's some serious overlapping sometimes, and as such slightly fewer Shreds than having a proper mangle bot, but I can't see how it would change that much.

I'd recommend strongly picking one of the two of you to be the manglebot and sticking with it. Both of your DPS should increase significantly.

My personal experience is that bosses do not have uniform armor, and as such rip does better on some fights than on others. That isn't proven though.

Please don't spread false rumors. All lvl80 bosses have the exact same amount of armor which is tested to be approximately 10643. Check http://elitistjerks.com/f31/t29453-c...gs_level_80_a/ for further info.

Rip ignores armor in any case, so it never does better or worse from boss to boss. The main variable is the debuffs present on the boss and your own armor penetration which changes the effectiveness of Shred. Nightcrawler's simulation shows the Shred Idol being superior under certain conditions. Unless you can prove with numbers that there is a logical fault in the simulation assumptions, challenging the conclusions with anecdotal evidence doesn't help anyone.

While you may be right that we'd see a DPS increase by deciding upon a mangle bot for every fight, it wouldn't seem it has got anything to do with my problem. My armor penetration is almost soft capped and between both trinket procs I'm at ~100% as often as it is currently possible. The Shred idol should by all means outperform the Rip idol, but it doesn't (according to limited observations).

I have long since lost the links, but I remember reading extensive testing by rogues and warriors in Naxx showing every boss to have the same armor. They were able to estimate the number and confirm the exact value using Maexnna. I haven't seen any such test done in Ulduar, but also have not seen any reason to believe they've changed that.

I just want to make sure about a question im having about ArPen versus Agility. Many times my guild will do 10mans and they'll be no warrior to sunder armor and at times they'll be no rogue (or rogue willing) to expose armor. Does the first post from NightCrowler that states Arpen>AGI (with mostly ulduar 25/10 gear) hold true only if Sunder armor or expose armor is up during the fight? I've read the post many times and I can't see where he states that it is understood that exposed armor or sunder armor x5 is a constant debuff or not.

Sorry if this was already posted I've done a read through and I can't seem to find an answer.

***I'm seeing he used the words "fully stacked raid" so I'll assume that means 5 sunders/or exposed armor throughout**

The way I understand it works is that ArPen takes away a percentage of the debuffed armor from a mob.

I think it's clearer if I just pull some numbers out of thin air to show. Let's say you have 25% ArPen.


Pretend Boss has 10000 armor; after ArPen, its armor to you is 7500.

Pretend Boss has 10000 armor and 5 stacks of pretend sunder removes 1000 armor for 9000 armor; after ArPen, its armor to you is .75*9000 = 6750.


So ArPen actually takes more armor away when the boss doesn't have any armor debuffs. Of course, if you have 100% ArPen, whether the boss has any armor debuffs won't matter at all to you.

As long as the target armor is greater than 7616.25, armor penetration is worth the same fractional damage increase regardless of the particular armor value.

Edit: on armor mitigated attacks.

Everyone is forgetting that a not insignificant portion of our damage ignores armor already, so its not quite that simple.

Apologies, I was using the wrong DR formula for armor.

100% armor penetration is a 50% gain in each case.

Boss has A armor after debuffs; assume A > C/2 (C = 15232.5). Damage reduction is

so fractional damage taken is
.

With 100% armor penetration, the armor is reduced to A-(A+C)/3. Damage reduction is then


so fractional damage taken is
.

This is an increase of 50% over the case with no armor penetration, and this holds for any armor value greater than C/2.

In general, the fraction of damage taken for A>C/2 is given by
,
where k is the armor penetration fraction. From this expression it's clear that the fractional value of armor penetration does not depend on armor (again as long as A>C/2).

Delete this. I'm stupid.

So right now where looking at:

Valorous Nightsong Headguard
Seed of Budding Carnage
Conqueror's Nightsong Shoulderpads
Drape of the Faceless General
Embrace of the Gladiator
Fluxing Energy Coils
Dreambinder
Idol of Worship (Rawr rates this as higher DPS)
Valorous Dreamwalker Handgrips
Soul-Devouring Cinch
Valorous Dreamwalker Legguards
Footpads of Silence
Cindershard Ring
Loop of the Agile
Blood of the Old God
Mjolnir Runestone
Gemmed entirely with Fractured Scarlet Ruby's (except for neck, which i placed Delicate Stormjewel in) to maximize potential dps?

Rawr gives that setup a 7698.807 dps rating.

- You are over the hit cap which makes certain items better then the ones you have here as BiS.
- Rawr's simulation is far from "real simulation", FeralByNight is more accurate and suggests that Idol of Ravenous Beast is better. I raid with ~700 ArP (I don't have Grim Toll / Mjolnir :-/) and agree with it.
- You have listed Valorous Nightsong Headguard instead of Conquerors Nightsong Headguard which is clearly better.
- I don't have any numbers to support this, but I'd go for 4T8. I think it was mentioned somewhere in this thread that it is approximately the same as 2T7. Anyway, I have tried it last night in ulduar25 and I love it. 1 cp SR lasting 22s? A lot more FBs that fit into rotation? Hell yes!

Hi all, after a small break from posting I'm back with some goodies.

First of all I've redone my simulation for cat druids with the Hard Modes discovered gears and the new 100% Arpen cap.

Code version: FeralbyNightv3_1_beta_w


Ulduar Best in Slot Gear
The following simulation is made with Ulduar best in slot gear. Actually PvP weapon is better but I don't like the idea of putting a PvP weapon in a PvE gear list.

Stats value

Stat	points		Error	dps value		Error
Armor Pen	30.3	+/-	-0.1	1.80	+/-	-0.01
Agility	25.0	+/-	0.1	1.49	+/-	0.01
Feral AP	12.0	+/-	-0.1	0.71	+/-	-0.00
Strengh	23.8	+/-	-0.1	1.41	+/-	-0.01
Crit	20.1	+/-	-0.1	1.19	+/-	-0.01
Hit	22.7	+/-	0.1	1.35	+/-	0.01
Expertise	22.7	+/-	0.1	1.35	+/-	0.01
Haste	21.1	+/-	-0.1	1.25	+/-	-0.01
AP	10.0	+/-	-0.1	0.59	+/-	-0.01

Slot	Item
Back	Drape of Faceless General
Chest	Embrace of the Gladiator
Feet	Footpads of Silence
Ring 1	Brann’s Signet Ring
Head	Conqueror's Nightsong Headguard
Legs	Conqueror's Nightsong Legguards
Hands	Conqueror's Nightsong Handgrips
Neck	Seed of Budding Carnage
Shoulders	Conqueror's Nightsong Shoulderpads
Trinket 1	Noble Deck: Agility
Trinket 2	Mjolnir Runestone
Weapon	Dark Edge of Deprivaty
Waist	Soul-Devouring Cinch
Ring 2	Cindershard Ring
Wrist	Fluxing Energy Coils
Consumable	Flask Flask of Relentless Rage + Rhino food
Professions	JC: BACK+CHEST, BS: GLOVES+WRIST

Gemmed with Arpen till the trinket soft cap (566 Arpen) and then gemmed with agility.

Simulation

Aggressive Stats after all the buffs (full stacked raid + Bloodlust used 1 time in a 5 min fight)
_______________________

Cat (crit% vs. a boss (-4.8% added)): 56.6%
Cat (AP): 10703
Expertise: 22
Hit: 7.2%
Armor Penetration: 46%
Haste (from equipment): 10%

Boss base Armor (before penetration and debuff): 10643
Boss modified Armor (after penetration and debuff): 4516
Boss Damage Reduction (after debuff and penetration): 22.9%
Weapon Speed Reduction % (Total without Bloodlust): 26%
Missed % (Total): 0.8%
Dodged %: 0.8%
Not normalized attacks parried by the boss: 0%
_____________________________________

DPS: 7878 +/- 1, (Min-Max variance: 31%, Lower dps sample: 6600, Upper dps sample: 9037) TPS: 5594


White cat: 2450, Yellow cat: 5428
White: 31.1% (SR: 0%)
Mangle: 0%, Avg dmg: nan, #/cycle: 0),
Shred: 32.8%, Avg dmg: 9870, #/cycle: 8.85),
RIP: 19.1%, Avg dmg: 40673, #/cycle: 1.25)
Rake: 10.2%, Avg dmg: 8879, #/cycle: 3.06),
Ferocious Bite: 6.7%, Avg dmg: 16528, #/cycle: 1.08, Average Energy Usage: 43)
Savage Roar uptime: 98.2%
Average Cycle Time: 33.7 sec.

The stats value are evaluated by subtraction, so if you are in naxx10, heroic gear, look at naxx25 table, if you are in BiS pre-Ulduar gear, look at Ulduar table. The step value to compute variable scaling was set to -143 stat points (I've chocen 143 because I need to pay attention to haste, In the simulation I've a temporal step of 0.01 seconds so the weapon speed it's a multiple of 0.01)

You can found dps value of all stats and item rankin here:
Gear Ranking

If you click on a number in the "score" collumn you will se an integer value, that value is exactly the dps value of that piece of gear


About Rake vs. Shred
Also at 100% arpen rake>shred due to lower energy cost.

About Arpen, Agility and 2T7 again

I've runned a simulation gemming with arpen instead of agility after the trinket soft cap. The dps is lower.
I've also runned a simulation with 2T7+2T8, at the BiS level the 2T7 gear choce is different than in naxx level gear and you must choce T7 legs and shoulders:

2T7+2T8	Item
Back	Drape of Faceless General
Chest	Embrace of the Gladiator
Feet	Footpads of Silence
Ring 1	Brann’s Signet Ring
Head	Conqueror's Nightsong Headguard
Legs	Valorous Dreamwalker Legguards
Hands	Conqueror's Nightsong Handgrips
Neck	Seed of Budding Carnage
Shoulders	Valorous Dreamwalker Shoulderpads
Trinket 1	Noble Deck: Agility
Trinket 2	Mjolnir Runestone
Weapon	Dark Edge of Deprivaty
Waist	Soul-Devouring Cinch
Ring 2	Cindershard Ring
Wrist	Fluxing Energy Coils
Consumables	Flask Flask of Relentless Rage + Rhino food
Professions	JC: BACK+CHEST, BS: GLOVES+WRIST

2T7+2T8 setup gives you 7811 DPS, So 4T8>2T7+2T8 by a small margin and it's mainly due to gear stats difference.


DPS dispersion

An Interesting analysis could be plotting dps dispersion for BiS gear:

%	Base DPS	Arpen	Agility	Feral AP	Strengh	Crit	Hit/Expertise	Haste AP
10%	7554	7309	7348	7355	7360	7389	7368	7382	7392
20%	7665	7417	7456	7460	7467	7496	7474	7487	7497
30%	7743	7493	7533	7540	7544	7574	7554	7567	7575
40%	7812	7557	7599	7608	7611	7642	7621	7634	7642
50%	7877	7619	7663	7672	7674	7706	7684	7696	7706
60%	7941	7681	7727	7733	7737	7770	7747	7760	7769
70%	8010	7748	7795	7801	7804	7838	7815	7829	7837
80%	8092	7827	7874	7880	7882	7919	7895	7911	7917
90%	8206	7937	7989	7992	7993	8032	8006	8021	8028

The following table show the dps decrease due to removing 1 point of each stat for dps sample under a certain percentile:

%	Arpen	Agility	Feral AP	Strengh	Crit	Hit/Expertise	Haste	AP
10%	1,71	1,44	0,70	1,36	1,15	1,30	1,20	0,57
20%	1,73	1,46	0,72	1,38	1,18	1,34	1,24	0,59
30%	1,75	1,47	0,71	1,39	1,18	1,32	1,23	0,59
40%	1,78	1,49	0,71	1,41	1,19	1,34	1,24	0,59
50%	1,80	1,50	0,72	1,42	1,20	1,35	1,27	0,60
60%	1,82	1,50	0,73	1,43	1,20	1,36	1,27	0,60
70%	1,83	1,50	0,73	1,44	1,20	1,36	1,27	0,60
80%	1,85	1,52	0,74	1,47	1,21	1,38	1,27	0,61
90%	1,88	1,52	0,75	1,49	1,22	1,40	1,29	0,62

And this table is stats point (fixing 10 for AP):

%	Arpen	Agility	Feral AP	Strengh	Crit	Hit/Expertise	Haste	AP
10%	30,2	25,4	12,3	24,0	20,4	23,0	21,2	10,0
20%	29,5	24,9	12,2	23,6	20,1	22,7	21,2	10,0
30%	29,8	25,0	12,1	23,7	20,1	22,5	21,0	10,0
40%	30,0	25,1	12,0	23,6	20,0	22,5	20,9	10,0
50%	30,2	25,0	12,0	23,7	20,0	22,6	21,2	10,0
60%	30,2	24,9	12,1	23,7	19,9	22,6	21,0	10,0
70%	30,3	24,9	12,1	23,8	19,9	22,5	20,9	10,0
80%	30,3	24,9	12,1	24,0	19,8	22,5	20,7	10,0
90%	30,2	24,4	12,0	23,9	19,6	22,5	20,8	10,0

Remembering that that value are given by sumbtraction we can see how Arpen and Hit/Expertise tend to "stabilize" dps while Agility/Crit tend to spred the gaussian to a wider range increasing tails size.









As you can see from the previous Image and from the dps-percentile table the difference for example between switching 143 Arpen and 143 agility is around 10 percentile, this mean that (doing numb math) if you're trying to compare agility vs. Arpen around 40% of times you will see an higher results for Agility and around 60% for Arpen. You can easly see how with a so great dispersion you can't see "in game" wich one is better.

I've also simulated the variable space for kitty dps. The simulation is done without raid buffs (except for istance heroism and "non stat buffs" like sunder, FFF and co), and without procs.
We can reduce the space of kitty dps to 5 dimensions: haste%, crit%, arpen%, hit/exp%, AP. The simulated dps value have an error of 5 dps (2 hours of simulations seems enough). I don't post conclusion about it but simply the output file. If someone wants to help me analazing it will be welcome:

Variables Space