Board Thread:Game discussion/@comment-4726614-20130404050044/@comment-6016076-20140330012239

Ok I have two theories:

Unlikely theory:

The bonus is removed by the reduction; there is some bug in the code where the bonus isn't applied but a +1 is added in its place.

Likely theory:

The game keeps track of armor to several decimal places, but the display value always rounds up. In addition, Garen's passive from W does not activate when it has been reduced. Using your example, we would have (46+30)*0.7=53.2. The W passive bonus is gone. The 53.2 is displayed in game rounded up to 54.

Applying this rule to your Shen example, though, would imply that he should display 185 armor since you came out with 184.1 in your calculation. However, you actually rounded up his armor initially when you summed his items, masteries, and runes from 187.6245 to 188. Using the true value 187.6245, Shen's total armor is 75 + 187.6245 = 262.6245, displayed in game as 263 (correct). 262.6245 * 0.7 = 183.8315, or rounded up in game to 184 (also correct).

So to me it makes the most sense that the game automatically rounds all true values up. Garen's W is bugged, or at least unclearly worded, for its passive bonus is removed when armor reduction is applied.  I long wondered about this myself, because if you open up the ability details on Garen's page here, you actually see a bullet point whose most likely interpretation means exactly this, suggesting that someone else may have done similar tests and discovered this detail as well. I'm glad I stumbled upon this discussion thread with some actual data. Somehow I missed it when I was wondering about this a few months ago.

To be fair, more tests should be run to definitively confirm this. Garen's armor after different stacks of Black Cleaver would probably be most ideal. This is a potentially game-changing bug though for his tankiness in late game battles where his W's passive can mean upwards of 40+ resistance lost. And if it's doing it to %reduction, who knows what %pen is doing (since we can't see it). We'd have to test how much damage someone with a LW does to him. It's a broken bug if the LW is also canceling his W's passive--that's a LOT of resists potentially lost!