ArP rating according to tooltip seems to have been lowered from 15.3 to something like 12.04.
Expertise change affects warriors aswell but the tooltip for this seems bugged (it shows the old value).

I have to admit that ArP rating and ArP effects (battle stance for ex) are really:
1) buggy.
2) completely different from before.

I start to think they changed the armor reduction formula.

I get 10647 armor for a boss target dummy (41.14% reduction) with 0 arp in berz
I get 39.09% reduction when in Battle with 0 arp but armor should then be 8,18% less and not 10%.

With arp on gear it is even worse

I did 3 tests with 0, 17.70% (218 rating) and 34.83% (429 rating) arp trying to keep ap close.
I used a spec with no dmg multipliers, no sunders, and I was alone on the dummy.
The Arp you get from Battle stance is not showed on charsheet (like mace spec), it's 10% reading the tooltip.
I got an approximate value for arp rating being 12.31640

	0 Arp zerk	0 Arp battle	17,7 Arp zerk	17,7 Arp battle	34,83 Arp zerk	34,83 Arp battle	0 Arp 5xSA zerk	0 Arp 5xSA battle
AP	3995	3995	3988	3988	3982	3982	3995	3995
Tooltip BT Dmg	1998	1998	1994	1994	1991	1991	1998	1998
Effective BT Dmg	1176	1217	1232	1277	1292	1341	1282	1325
Reduction %	41,1411%	39,0891%	38,2146%	35,9579%	35,1080%	32,6469%	35,8358%	33,6837%
Expected Armor	10647,21	9775,33	9421,4	8552,63	8241,11	7383,39	8507,39	7736,96
Effective Reduction		8,1870%	11,5112%	19,6710%	22,5969%	30,6529%	20,0959%	27,3320%

I hope you can use some of these data to find the interactions, because currently those values aren't equal to armor*0.9 nor armor/1.1.

Unless armor formula has slightly been changed, but that would require a target with known armor.

Landsoul and I did our same testing again, and found that arpen is still working oddly, as before. Let me know if anyone wants our data/calculations, but it's basically identical before, except with 25% more arpen from rating. The effects are still wonky.

Is it just oldRating*0.75 or am I getting something wrong there?

Its currently bugged that instead of multiplying by 1.25, they divided by 1.25.

Battle Stance = 10/1.25 = 8 (yes, its double bugged since its not even a rating, lol)

218 ArP = 218/15.39529991/1.25 = 11.32813268

429 Arp = 429/15.39529991/1.25 = 22.29251798

Sunder is not bugged and gives 20%.

Combinations seems to be additive for all of your data except 0 Arp 5xSA battle, which is somewhere between additive and multiplicative.

Boss armor seems closer to 10610, using the above calculated reductions to calculate "Effective BT Dmg" matches your data within error of 1 damage.

Thanks for your data

It'll be interesting to see if the bugged ArP was fixed in the latest PTR build.

Also, has anyone run similar tests with a Blood DK and the effects of Blood Gorged? I am curious if it has the same problem as battle stance. If not, I'll be on PTR tonight with a Blood DK and can collect some numbers, so hopefully I'll have a decent amount of data to contribute to the efforts.

@Hellord-If possible can you post the spec and a chardev sheet? It would be good to know exactly what talents and gear you were using for the tests.

@Whitetooth-Keep up the good work, I really appreciate your efforts with rating buster and contributions thus far to the community as a whole.

During the most recent tests I did on PTR I found out that ArP rating is working with the same rating to% as on live.
I managed to equip 54% Arp (674 rating) but the armor reduction seemed to be only 44sh% (more or less same conversion on live).
Battle stance seems to be around 8.2% armor ignored rather than 10%, and when you stacks debuffs the values changes aswell.

The expected armor using multiplicative or additive method is never close enough to results and differs much the more debuffs you have.

I made tests with a spec without multipliers, using BT with 3607AP (1804 BT damage on tooltip) with 674ArP rating from gear (54.72% on tooltip)

http://talent.mmo-champion.com/?warr...0&version=9684

I am still clueless of what happens behind the stacks, either it is due to a bug or a rating conversion change different from what tooltip says.

I prefer to include an xls this time so everyone can check if I did any mistake.

http://www.hellord.com/wow/PTR_Arp_stack.zip

TLDR: At first glance Battle stance is only providing 8%sh armor penetration and ArP rating on gear seems not 25% better than live.

I did few more tests on this build and I would like to share the results.

Actually I'm inclined to think that they "tend" to stack additively, but there are still some strange behaviours that I can't understand.

All tests were done with a spec without multipliers, using Bloodthirst with 4295AP for 0arp gear and 4270AP for 356arp gear. Since Bloodthirst have a 1 damage difference between max and min I always waited to get both values and averaged.

I examined the 2 armor debuffs (Sunder and FF) plus a spec buff plus gear buffs, separately and together.

Sunder 5 stacks alone is consistently around 20% reduction, the difference is due to roundings most probably.
FF alone is 5% reduction, roundings skew this value aswell by a few decimals.
So the debuffs seems to work correctly.

Battle stance should give 10% Arp, but used alone seems to only give between 8.02% to 8.2%. Since the value is quite much lower than expected I'll have to make more tests with full raidbuffs to have a better BT damage and closer results.

356 Arp rating shows (from tooltip) 28.90% armor penetration, but from tests it seems to only grant around 23.2%-23.4%. This value is very close to the ArP rating conversion on live server (so 25% less than tooltip).

So after seeing that Arp from gear and battle stance grants different buffs from what tooltips stated, I used those values to check how they stack. ~8% for BS and ~23.3% for Gear arp.

I used an arbitrary boss armor value of 10650 that should be quite close to what is real (values from 10645 to 10655 where inline with expectation in most cases).

Stacking FF and BattleStance
Stacks additively (8+5) (the difference from expected is around 0.03%) giving ~13% armor penetration

Stacking Gear, FF and BattleStance
Stacks additively (23+8+5) (difference from expected around 0.5%) giving ~36% armor penetration

Stacking Sunderx5 and BattleStance
Stacks additively (20+8) (difference from expected around 0.54%) giving ~27.5% armor penetration

Stacking Gear and FF
Stacks additively (23+5) (difference from expected around 0.3%) giving ~28% armor penetration

Stacking Gear and BattleStance
Stacks additively (23+8) (difference from expected around 0.15%) giving ~31.6% armor penetration

These values are quite close to those I would expect considering BattleStance an 8% buff in ArP and gear arp with live rating conversion (25% less than expected).
What I noticed is that when you factor in sunder armor the values skew by much more.

Stacking FF and Sunderx5
Stacks additively (20+5) (difference from expected around 1%) giving ~24% armor penetration

Stacking FF, Sunder and BattleStance
Stacks additively (20+8+5) (difference from expected around 1.65%) giving ~31.4% armor penetration

Stacking Gear and Sunderx5
Stacks additively (23+20) (difference from expected around 1.9%) giving ~41.5% armor penetration

Stacking Gear, FF and Sunderx5
Stacks additively (23+5+20) (difference from expected around 3.2%) giving ~45.2% armor penetration

Stacking Gear, FF and BattleStance
Stacks additively (23+8+20) (difference from expected around 2.5%) giving ~49% armor penetration

Stacking Gear, FF, BattleStance and Sunderx5
Stacks additively (23+5+8+20) (difference from expected around 4%) giving ~52.5% armor penetration

Using proper values for gear and Battlestance makes this values shift by quite much, so assuming a bug I preferred to use parsed values.

After all seems that sunder is the one creating problems when stacking since it varies from 0.54% to 4% difference from expected.
I compiled a sheet with all results and data with comparisons, hope somebody can use them and maybe complete them with better data.

PTR_Arp_stack_b9722.zip

I can't look at your spreadsheet right now, but the numbers that you listed work better if we assume that there are 2 kinds of armor penetration. The first is Sunder and FF, which stack multiplicatively. The second is Battle Stance and gear, which also stack multiplicatively. You then stack those 2 sources of armor pen additively. However, this only applies if Sunder is present. If there's no Sunder the numbers are much closer if Battle Stance and gear are additive, along with FF.



Stacking FF and BattleStance
Stacks additively (8+5) (the difference from expected is around 0.03%) giving ~13% armor penetration
Matches my assumptions

Stacking Gear, FF and BattleStance
Stacks additively (23+8+5) (difference from expected around 0.5%) giving ~36% armor penetration
My assumptions predict 34.16% armor penetration, which is less accurate than what you came up with

Stacking Sunderx5 and BattleStance
Stacks additively (20+8) (difference from expected around 0.54%) giving ~27.5% armor penetration
Matches my assumptions

Stacking Gear and FF
Stacks additively (23+5) (difference from expected around 0.3%) giving ~28% armor penetration
Matches my assumptions

Stacking Gear and BattleStance
Stacks additively (23+8) (difference from expected around 0.15%) giving ~31.6% armor penetration
My assumptions predict 29.16% armor penetration, far less accurate than if they are additive

Stacking FF and Sunderx5
Stacks additively (20+5) (difference from expected around 1%) giving ~24% armor penetration
FF and Sunder stacking multiplicatively matches your results exactly, and is what others such as Astrylian have reported

Stacking FF, Sunder and BattleStance
Stacks additively (20+8+5) (difference from expected around 1.65%) giving ~31.4% armor penetration
My assumptions predict 32% armor penetration, so only about 0.6% off

Stacking Gear and Sunderx5
Stacks additively (23+20) (difference from expected around 1.9%) giving ~41.5% armor penetration
Matches my assumptions

Stacking Gear, FF and Sunderx5
Stacks additively (23+5+20) (difference from expected around 3.2%) giving ~45.2% armor penetration
My assumptions would predict 47% armor penetration, not too accurate but slightly closer than assuming they are all additive

Stacking Gear, Sunderx5 and BattleStance
Stacks additively (23+8+20) (difference from expected around 2.5%) giving ~49% armor penetration
My assumptions predict 49.16% armor penetration, matching your results

Stacking Gear, FF, BattleStance and Sunderx5
Stacks additively (23+5+8+20) (difference from expected around 4%) giving ~52.5% armor penetration
My assumptions predict 53.16% armor penetration, somewhat closer to your results than assuming everything is additive

Your point is completely valid.

It seems a rough formula can be 1-((1-sunder)*(1-ff) -bs -gear).

I am gonna try mace spec to see if there is a 25% damage reduction there aswell (as it is for battle stance).

EDIT: Also mace spec seems to only remove 12%, so 25% less than expected; same happens to battlestance and gear and if anybody can check Blood Gorged (DK blood) and combat rogue mace spec we can probably state that these are semi-conclusive tests.

I can't post on US forums, so it would be nice if somebody else can give feedback on this bug.

I was also wondering if there have been a slight change in armorToDR formula, but I didnt make any test on a target with known armor.