Alo.
I will try to explain it with a simple example.
Consider that we have 3 types, A, B, and C. The attacker can deal a total of 600 damage, while the defender is Good on A, Normal on B, and Bad on C, in order of course to have a Normal average, i.e. "Good + Normal + Bad = Normal". The damage resistance of these types for a full WM 235+10 archer armor are 255.43, 208.03, and 160.62, respectively. At this point I want to stress out that the above-mentioned values are real and correct, and have been calculated assuming real armor.
In the following figure you can see 4 distributions of the offender's damage. In the Red case he distributed uniformly, while in the remaining 3 cases he just picked one of the types. Results show that the uniform distribution (Red) is equal to picking damage type B (Green), because of the Normal average on armor. Picking A (Blue) yields to lower damage, while picking C (Magenta) to higher one.
The problem of course is that you do not know what the defender's resists are. So what I am saying here is to uniformly distribute it along all types, since this choice will result to the
minimum volatility of your damage along different opponents and time. I am not saying that it will be the max you can do! I am just saying that you will have the less dispersion of damage, with minimum max/min damage surprises over time.
If you think of it really deeply, you will find out that there is no point on this whole discussion actually, since in the very very very longrun and always statistically speaking, you will find opponents with Good or Bad on one type, if you really select this one and only type. But the volatility of damage will be surely larger and more frequent over opponents and time. And this is something I personally do not like.
Best,
Dr. LawZ
PS. I am so happy when we discuss for such things and not for "who has it bigger", as I can see in some other topics.