Combat Ratings at level 85 (Cataclysm) - Page 26

acceleratum · 03/28/09, 2:37 AM

what about defense vs resilience for crit immunity?

As a warrior tank im interested in this topic, does defense have a steep diminishing return? I find usefull to add some resilience in order to reach crit immunity without wasting every single gem/enchant into defense.
Im lookin to build a good allround tanking gear but the defense cap leaves alot of good options out, with some resilience in it, I believe a better overall gear could be achieved.

Does anyone got numbers on this two stats related to crit immunity? I couldnt find it around the forums.

Rallik · 03/28/09, 4:18 PM

I've been working at figuring out how arpen/debuffs stack today and have a theory for a formula:

$\frac{A}{A+B(1-(x+y)+\frac{B}{A+B}xy)}$

where:
A = the half armor value at 80, 15232.5
B = base boss armor, 10643(the value I found from testing)
x = sum of arpen buffs(mace spec/battle stance/gear arpen)
y = product of arpen debuffs(sunder/ff) (1-.95*.8)

Currently arpen buffs are not working as specified, and are all .81 of what they should be(note: not .8 like you'd expect from 1/1.25). I.e. Battle stance = 8.1% reduction, mace spec = 12.15%, and 20% arpen from gear(per the tooltip) = 16.2%. The nice thing is that they do stack together additively. So, if you had Mace spec/Bstance/20% arpen, x would be .81*(.15+.1+.2) = .3645

Sunder/FF on the other hand stack together multiplicatively, so 24% reduction when both are applied.

The debuffs and buffs together however do not interact as I'd expect at all. The buffs are summed together, but then do not stack multiplicatively or additively with the debuffs, but somewhere in between. After fiddling with the numbers for a long time and trying to discover a relationship, the ratio in between seems to be A/(A+B), the armor reduction for a boss at full armor oddly enough.

To hopefully be a little more clear.. if you had 10% arpen from gear(ignore the .81 bug for now) and Sunder/FF up.

expected armor reduction from arpen alone if fully multiplicative = .1 * (10643*(1-.24)) = 808.868
expected armor reduction from arpen alone if fully additive = .1*10643 = 1064.3

Actual armor reduction from arpen = 808.868 + (A/(A+B))(1064.3-808.868) ~= 959.24, which puts the final armor value at 10643*.76-959.24 ~= 7129.44, which becomes (15232.5/(7129.44+15232.5)) = ~= 68.11% damage, which is the result you'll get if you plug all those numbers into the equation listed above.

Does this check out with everyone else's testing?

• Melthu · 03/28/09, 9:52 PM

Your equation matches hellord's data within 0.1% of his observed armor reduction, which can probably be chalked up to rounding. If you came up with it using your own data I'd say it's pretty likely that this is the equation being used in-game.

Rallik · 03/28/09, 10:44 PM

Originally Posted by Melthu

Your equation matches hellord's data within 0.1% of his observed armor reduction, which can probably be chalked up to rounding. If you came up with it using your own data I'd say it's pretty likely that this is the equation being used in-game.

Yep, it was derived from my own data, and matched every combination I tested, good to hear it works for others' as well.

Mijae · 03/30/09, 2:09 PM

Originally Posted by Rallik

Currently arpen buffs are not working as specified, and are all .81 of what they should be(note: not .8 like you'd expect from 1/1.25). I.e. Battle stance = 8.1% reduction, mace spec = 12.15%, and 20% arpen from gear(per the tooltip) = 16.2%. The nice thing is that they do stack together additively. So, if you had Mace spec/Bstance/20% arpen, x would be .81*(.15+.1+.2) = .3645

Sunder/FF on the other hand stack together multiplicatively, so 24% reduction when both are applied.

First, great work.

However, you state the 0.81 modifier is not equal to 1/1.25. Do your results show that the 1.25 multiplier is in fact applying to armor pen rating? Is it already present in the tooltip value? Could there be some combination of this new multiplier affecting rating, but not talents that makes it appear not exactly 0.8?

If no ArP buffs other than rating:

x = rating / 15.39529991 * 1.25 ? * 0.81 ?

Exemplar · 03/30/09, 2:54 PM

What I'm getting is that ArmPen is still "broken".

Either it's effectively unchanged, tooltip buffed by 1.25 but real-tests providing 80% of tooltip, or effectively (* 1.25 / 1.25). Net approx unchanged from live. Or instead of being 1.25 times as good (multiplied by 1.25) it's .8 times as good (divided by 1.25). In short, ArmPen is worse on PTR than on live. I'm personally not clear on which situation just from reading.

If either is indeed the case, is there a bug thread opened on the topic in addition to any in-game PTR reporting which may have occurred? Neither is matching what Blizzard has advertised or seems to intend.

Rallik · 03/30/09, 3:54 PM

The tooltip value alone is showing the 25% increase in arpen. If you had 5% arpen from rating on live, you'd see 6.25%(in the tooltip) on the PTR. 81% of that tooltip value is what is actually being applied however. So, if you had 5% arpen on live, you'd have 5*1.25*.81 = 5.0625% on the PTR(rating / 15.39529991 * 1.25 * 0.81). It seems very strange that it is not precisely 80% which would be expected, but I double-checked with battle stance/mace spec/gear in various combinations many times, and always saw the same 81% reduction. So yeah, it's still broken, and the very odd interaction between arpen buffs and debuffs seems like a malfunctioning mechanic that can't be explained well either. Hopefully the build going up soon finally addresses this bug.

Etaiu · 04/02/09, 9:21 PM

Originally Posted by Selmarix

Unfortunately that requires a lot of testing that is difficult to do (because of no direct way to retrieve miss values after DR).

So someone would have to get an equal level mob to attack a lot of times (without dieing) and then the same for different amounts of defense equipment.

That is far more time consuming and difficult than the parry and dodge data collected in this thread which is probably the reason why no one has done it yet.

The other requirement for such a test is to have so much avoidance that you are never struck by normal hits, ever. It's not an issue of the amount of data gathered.

nightcrowler · 04/03/09, 9:08 AM

I'm not sure that ArPen is bugged. Actually we are assuming that armor penetration doesn't suffer diminushing return but, if it's not the case? I'm under the impression that it's not a bug or tooltip bug but that actually arpen cap at 100% armor penetration but looking at the formula if 100% arpen really gives 100% armor pen we will obtain a negative value for boss damage reduction if we cap ArPen. If we now put the 0.81 penality the minimum damage reduction will be 2% an ok value it seems. Basically what I want to say is that the 0.81 shouldn't be put into ArPen formula but into damage reduction formula while taking ArPen into account. In that way they really have increased ArPen but they also have changed Armor formula to avoid negative reduction values.

hellord · 04/03/09, 10:24 AM

Originally Posted by nightcrowler

I'm not sure that ArPen is bugged. Actually we are assuming that armor penetration doesn't suffer diminushing return but, if it's not the case? I'm under the impression that it's not a bug or tooltip bug but that actually arpen cap at 100% armor penetration but looking at the formula if 100% arpen really gives 100% armor pen we will obtain a negative value for boss damage reduction if we cap ArPen. If we now put the 0.81 penality the minimum damage reduction will be 2% an ok value it seems. Basically what I want to say is that the 0.81 shouldn't be put into ArPen formula but into damage reduction formula while taking ArPen into account. In that way they really have increased ArPen but they also have changed Armor formula to avoid negative reduction values.

Arp doesnt seem to suffer diminishing returns. I tested quite a lot of combination of debuffs and gear (up to 60% arp from gear) and the relative contribution due to gear is always a flat 80%sh. What diminishes is the contribution of FF and Sunder at higher level of arp from gear.

Also due to the multiplicative nature of stacking debuffs, there is no way to reach more than 100% arp and get a negative value.
Even at 99% arp from gear if you apply both sunder and ff the boss is still left with 0.76% armor and the only way to make it 0 is having 100% arp. Overcapping it will just be a waste like it is for other cappable stats like hit and exp (or crit) and if they want to control this cap they just need to itemize gear properly.

There is currently no explaination for arp reduction. The 0.81 value can even "work as intended" but it's meaningless saying that battlestance reduces armor by 10% since it doesnt happen, nor saying that arp rating got a 25% buff sicne tooltip shows that but the game just re-reduce it to something like actual live value+0.8%.
The real changes are debuffs contribution (changed to relative) and gear/buffs stacking that is now additive. So while buffs won't change the value of your arp on gear (they only lowers the cap), both debuffs applied will reduce it's value by 25% (0.76 factor if you apply both FF and sunder.

Putting a DR on arp could be a way to reduce it's scaling, that is actually increasing, but honestly i didnt read anywhere a blue sayin anything about it.

EDIT: I did few tests today and BattleStance still gives 8.1% arp, so nothing has changed yet

nightcrowler · 04/03/09, 10:46 AM

I don't get it Hellorco. If I apply Rallik formula without 0.81 suppression I can have 100% armor pen from gear (I'm a druid so I don't have talents) and another 24% from debuff (FFF+Sunder).
In formula:

x=1
y= 1- 0.95*0.80 = 0.24
A = 15232.5
B = 10643

15232.5/(15232.5+10643*(1-(1+0.24)+(10643/(10643+ 15232.5))*(1*0.24))) = 1.1095 = 110.95% is more than 100%

if now I apply the formula with 0.81 suppression:

15232.5/(15232.5+10643*(1-(1*0.81+0.24)+(10643/(10643+ 15232.5))*(1*0.81*0.24))) = 0.9795 = 97.95% is less than 100%

this is what i wanted to say: without 0.81 suppression you could put the boss at "negative armor" (nonsense) so probably there is a "new armor formula" and 0.81 is simply a factor similar but indipendent from the 1/1.25 of the new arpen conversion rating.

Exemplar · 04/03/09, 12:05 PM

What Hellorco is saying is that if you get enough ArmPen that the Armour rating of your target would go below 0, nothing happens. The extra ArmPen is wasted.

This is identical to how Hit and Expertise function. If you get to 0 misses with Hit Rating, extra Hit rating does nothing. If you get to 0 dodges and parries with Expertise, extra Expertise does nothing.

You would reach 100% Armor Reduction lower than 100% ArmPen. End of story.

dysent · 04/03/09, 12:30 PM

Regarding .81 suppression on ArP:

If you do the math based on .81 suppression, ArP on PTR is 1.0125 * live
Tooltip is 1.25*live.

So: if you take the way it works on live, and just increase by 1.25% instead of 1.25, but increase the tooltip by the full 1.25, you get the strange value it has now.

Occam's Razor: Sounds like a coder flubbed their decimal point

[edit: posted this on blizz PTR forum using Hellord's data (hope you don't mind - not trying to take credit for your effort - just had lots of time bored@work to build a [bug] post and want to see this crap fixed) This time Deso chimed in with a 'to be examined', which is something I didn't see in the other ArP bug threads; possibly because they got muddied down by the sunder/FF interaction, which isn't actually a bug and confuses the data. Hopefully the extra visibility helped.]

• Melthu · 04/03/09, 2:45 PM

The thing is, the 81% effectiveness bug isn't just happening to ArPen from gear. Penetration from Battle Stance and Mace spec are also affected by it, so something else is going on.

Exemplar · 04/03/09, 3:26 PM

Originally Posted by dysent

[edit: posted this on blizz PTR forum using Hellord's data (hope you don't mind - not trying to take credit for your effort - just had lots of time bored@work to build a [bug] post and want to see this crap fixed) This time Deso chimed in with a 'to be examined', which is something I didn't see in the other ArP bug threads; possibly because they got muddied down by the sunder/FF interaction, which isn't actually a bug and confuses the data. Hopefully the extra visibility helped.]

Bold is my highlight - that's a great sign. Hopefully in a new push this will be corrected (125% not 1.25%) and it'll be easy to verify working as expected.