Go to the Delvers and Denizens Index
Dan Collins at Delta's D&D Hotspot recently did a statistical deep dive into the weapon vs armor table in the 1e AD&D PHB. Essentially the numbers come from the Chainmail man to man melee table, but Gygax did his math wrong. This is why the numbers in the PHB don't even remotely reflect historical reality. The numbers in Chainmail aren't exactly historically precise either, but at least they feel like they could be reasonable.
In my last effort at putting together houserules for OSE I considered using a weapon vs armor table, but since the one in AD&D is so ridiculous I tried to make my own. To do this I thought about restricting AC just to specific armors as originally intended and have dex add to hp instead of con. However, in the end I decided against it because if AC equaled actual armor types then I would have to redo the AC for every single monster, which I wasn't about to do. So AC is an abstract concept representing how difficult it is, or how much effort is required, to actually cause a significant injury. I tried assigning specific +or- values to hit for each weapon vs armor types, e.g. giving +2 to hit chain for maces to represent the fact that the chain doesn't protect against bludgeoning weapons, but the padding underneath would. I found it too fiddly though, and I ended up giving certain weapons Target Specialization(granting advantage or disadvantage) vs certain armor types, such as chain, or padded, or plate, and this bonus would apply to monsters with a similar type of 'natural armor' like dragon scales, etc. So this is what I came up with:
So spurred on by Dan's analysis and by
this post with a further conversion of the chainmail numbers, I decided to do my own take.
First, since chainmail used a 2d6 as a basis we can assume that a 7 to kill is average. So going from there I extrapolated the probabilities of getting below or above seven and converted that to a bonus or penalty, numbers above 7 result in some -%, and numbers below seven result in some positive %:
then I took the numbers needed to roll a hit in Chainmail, resulting in a kill, and converted the 2d6 number to a % bonus or penalty. For the missile weapons I took an average value of all three ranges:
From this it's simply a matter of dividing by five to convert to a d20 to hit roll as +or-1 is the same as +or-5%. I ignored the values for shields as I have my own rules for shields and their use is very situational, as opposed to armor which is always there. So this is what the Chainmail values to kill would be converted as adjustments to hit:
Some of the numbers make sense to me, like the huge penalties to bows vs plate; its not really any of the numbers that bother me, but comparing them to each other. What's the difference bewteen a horse bow and a composite bow? I'm assuming a horse bow implies a bow used by the different steppe cultures, but all those bows were composite, so why have two listings? And why is the 2-handed sword better at killing someone in armor than the lance? That is so ridiculous. and why are the polearm and halbard so drastically different? did Gary not know that a halbard is a polearm? And the Spear, I mean come on. And the drastic differences between a mace, morningstar, and flail; why would weapons that are so similar have such huge differences in effectiveness?
So, I think we can conclude that Gary set a few of these numbers based on what he knew about history and pulled the rest out of his ass. I think I'm going to stick with advantage/disadvantage as it better represents the vicissitudes of combat, rather than try to work this table so it represents reality.
