Talk:Armor/@comment-27371060-20151209110346/@comment-5723439-20160125161307

No, it doesn't. It illustrates that the marginal gain of EHP per unit of armor per unit of armor for any given amount of hp is always decreasing.

All that needs to be shown is the following:

The first partial of EHP with respect to armor depends on hp and the first partial with respect to hp depends on armor. The reasoning why this method is superior to looking at a third partial is because hp and armor are not currencies; gold is a currency. You want to maximize the EHP from the amount of gold you have. If HP were 10x more expensive, it doesn't matter if +150 armor and +1000 hp places you on the optimal vector in the EHP surface. The vector costs 25k gold, and you can't afford that. In fact, using the 10k gold you have to buy full armor nets you more EHP than trying to follow the vector. We care about gold and what goldcan buy you, not what's theoretical.

But even more important, publishing the % gain is misdirection from what I can tell. It would imply to the masses that the absolute gold efficiency of EHP per armor goes down as you get more of it. That's simply not true. It remains constant. The gold efficiency doesn't even go down. What does go down is potential gold efficiency, but that's not because armor is getting less effective; it's because HP is becoming more effective. The above publication doesn't address that caveat at all.