An Elegant Solution to ArPen Scaling

#0 - May 15, 2009, 9:58 p.m.
Blizzard Post
I'll begin with a description of how ArPen currently works, and the issues that arise with its implementation.

The first thing you have to know is the armor of a target after debuffs have been applied. I'll be using the example of a boss level mob, which has a base armor value of 10643. Applying sunder and FF to the target reduce it's armor by 20% and 5%, multiplicatively. This gives a final armor value of 10643*.8*.95 = 8088.68.

Next, you determine the ArPen "cap" which limits the effectiveness of ArPen against high-armored targets. This is mostly necessary for PvP reasons, as 100% armor penetration would otherwise reduce any target, regardless of their armor, to 0. This is calculated as: (15232.5 + ArmorAfterDebuffs)/3. In the case of a boss mob, this cap is (15232.5+8088.68)/3 = 7773.73. If this value is higher than the armor of the target, the armor of the target is instead used. I.e. a clothie with 2000 armor will have a cap of 2000 rather than (15232.5+2000)/3 = 5744.17.

Next, you multiply the sum of ArPen rating and character buffs by the calculated "cap" to determine the actual amount of armor you are reducing. With Battle Stance(10%), Mace Spec(15%), and 30% ArPen from gear, this would be (.1+.15+.3)*7773.73 = 4275.55 armor reduction.

Finally, you plug this all into this formula to determine your damage modifier on the target: 15232.5/(15232.5 + Armor - ArPenReduction). Using the continuing example, this would be 15232.5/(15232.5 + 8088.68 - 4275.55) = 79.98% damage dealt.

This system has the issue that has been brought up recently, that there is no cap on armor penetration, and that armor penetration scales exponentially, as it has an inverse relationship. For example, at 300% arpen, your ArPen reduction becomes 3*(15232.5 + 8088.68)/3 which clearly equals 15232.5+8088.68. If you plug that into the formula, you get a division by 0... 15232.5/(15232.5 + 8088.68 - 15232.5 - 8088.68). This is why it scales up to infinite damage(although it is relatively unachievable for now).

AN ELEGANT SOLUTION THAT SOLVES EVERYTHING:

Rather than reducing the armor value itself, which has an inverse relationship with damage dealt, resulting in the exponential scaling and the issues that result, it makes more sense to turn the equation around and rather than reducing the armor, just increase a character's ability to deal damage through a target's armor. To clarify, this is much akin to how Haste effects increase your attacks per second rather than reducing your seconds per attack(which would have an inverse relationship and exponential scaling up to 100% haste).

To change the damage formula to implement this, you merely need to shift the "cap" into the numerator rather than the demoninator. That is: (15232.5 + ArPenReduction)/(15232.5 + Armor). This eliminates the exponential scaling of ArPen by turning it into a linear function, and maintains the current functionality of the system, in that arpen will have relatively little effect against cloth, while also not being devastating to highly armored targets. It also does this without needing a hard cap, since it will never approach infinite damage without... infinite ArPen.

Here's a screenshot of a graph to demonstrate this effect on a Sundered/FFed Boss mob. The lines intersect at the point where armor is reduced to 0, which is 104.05% on a boss mob. You can see that this change would cause ArPen to be nearly equivalent to its strength now up until you reach 0 armor/100% damage dealt, at which point it will continue to scale linearly rather than increasingly exponentially.

http://i163.photobucket.com/albums/t313/Sabotage101/ArPen.jpg

Please consider this as a solution to the ArPen problem. Eliminating exponential scaling without the need to introduce a hard cap and also doing away with all the "So do I gem arpen after 30%??" questions due to the nature of the formula would be well received by the community. Balancing the strength of ArPen rating with a linear formula will also be far easier when there's no need to take into account the runaway effect that exponential scaling could lead to in the future.
#31 - May 17, 2009, 8:34 p.m.
Blizzard Post
Q u o t e:
Another brilliant post with an ideal solution that will not be used, or even considered because "Blizz knows best"


Don't ruin a good post with trolling.

Also please don't bump. If players are interested, they will respond. We read everything, even stuff on the last page.