The Big Mistake in Weapon vs. Armor Adjustments

Players of 1E AD&D duel with the most heavy-weight table in the PHB (p. 38): the "Weapon Types, General Data, and 'To Hit' Adjustment" table, which includes Armor Class Adjustments, intended to recreate the matchups of certain weapons versus certain classes of armor. 

AD&D 1E PHB Weapon TYpes Table

It's among the more complicated things in the game. On the one hand, they're not listed as variants or optional rules; and they're ingrained to the DMG example of combat -- so many 1E players do pound these tables into their games, determined to faithfully use them no matter how awkward they are. On the other hand, at least as many players of the game overlooked them, and Gary Gygax is even on record as saying the same thing. Let's document a few quotes:

I did not use psionics, generally ignored weapons vs. armor type and weapon speed. – Gary Gygax, ENWorld Q&A, 24th January, 2003

There is often player pressure to add complexities and complications to rules and systems, such additions being urged in areas that the players like and believe to be critical to enjoyment of the game. I did that for some writing in OAD&D and regretted it considerably thereafter – mainly weapons vs. armor types and psionics. – Gary Gygax, ENWorld Q&A, 24th July, 2003

In all, I included the details because of insistance of some avid palyers that were in touch with me, regretted listening to them, for the RPG is not suited to combat simulation... As I noted above, we never used the weapons vs. armor type adjustments. – Gary Gygax, ENWorld Q&A, 7th September, 2005

He elsewhere says the same thing for the Space Required value given to each weapon; that is, Gygax wants to entirely ignore everything on the weapons table taking up all of AD&D PHB p. 38. (And his son Luke Gygax reiterated the same thesis when we interviewed on him on the Wandering DMs, the day after I wrote this.) But if that's the case, it begs the question, where exactly did these numbers come from? Were they ever playtested? The questioner on ENWorld Sep-7 2005 asks this question, but Gary lightly dodges it.

As usual, the answer is given more clearly if we take a step back to the Original D&D texts. Essentially these same modifiers first show up in an earlier table in OD&D Supplement-I, Greyhawk, p. 13-14. (I'll show a recreation of that chart a bit later.) Right before this table Gary writes the source for these numbers: 

For those who wish to include weapon types in the determination of hit probabilities the following matrix drawn from the “Hand-To-Hand Combat” section of CHAINMAIL is offered. If this system is used it is suggested that the separate damage by weapon type and monster type also he employed.

Okay, so we need to take one more step back to Chainmail and to the actual origin of these numbers. Here's the earliest expression of that work to which I have access, from Chainmail 2nd Edition (p. 37):


Notice the presentation here is different; the numbers in Chainmail aren't modifiers to D&D-style attack rolls; they're final target numbers for the attack. That is, this system effectively assumes that all combatants are the same level and strength ("normal men", you might say), with the basic factor in this system being only the weapon and armor employed on each side. (They're also target numbers on a 2d6 roll, instead of D&D's d20 roll, but that's not the most salient thing.) So how do those different systems correlate? Note that Gygax is a bit cagey on this point -- he says, "the following matrix drawn from the 'Hand-To-Hand Combat' section of CHAINMAIL", but he doesn't say exactly how it was drawn. And here's where the math comes in, which allows us to pierce the veil, and see exactly what he did.

I've tried a few things, and I won't present all the failed investigative paths on my part, but here's the recreation that works the best. Gary first takes the average of all the target numbers in the Chainmail table; we'll call this the "base to hit", and it works out to about 7.8, or simply 8 if you round it off. Next, he just subtracts every number in the table from this base value of 8 to see what the effective modifier is in each weapon-vs-armor combination. That's it; pretty simple, really.

You can see these computed differences compared to the numbers that appear in Greyhawk below (and afterward, mostly repeated in AD&D). The numbers match very closely; this is clearly what Gary did.

Chainmail Man-to-Man Melee Chart and Conversion

Greyhawk Weapon Adjustments and Errors

More detail on exactly how closely these match: As shown in the last table above, most of the entries in Greyhawk vary from our computation by either 0 or 1 pips. The sum of all the differences in the entire table from our formula-based conversion is only 30 pips total. Hypothetically, if you use any number other than "8" as the basis (the average of all the targets in the Chainmail table), then you get much more divergent differences (far bigger sum of absolute errors).

In the number of cases where our formula differs from what appears in Greyhawk, clearly Gary was bothered by the larger penalties, and shaved them down to something less onerous. In particular, it looks like he was especially troubled by the line for "Spear"; with our formula it mostly has penalties all down the line, up to a hefty "−4". Gary joins many of us amateur historians in recognizing the spear as a dominant force on the real-world battlefield, and in response, dials down the penalties by about half. That's the only row in our "Absolute Error" table where the adjustments are as much as 2 points off our calculated formula; and we empathize with that. 

But wait. 

If the Spear made sense in Chainmail (presuming it did), then why is the translated version giving us such a headache in Greyhawk, anyway?

Can you see the gaping flaw in the conversion method?

It's huge, and it's obvious, but I didn't see it for decades, and I've never seen anyone else point it out.

I'll give you a minute to think about it.







It fails to recognize the natural protective value of the armor included in the Chainmail table.

Even if your "base to hit" in Chainmail was 8 versus Leather, say, it shouldn't also be 8 for Chain and 8 for Plate across the same row. It should be adjusted by some amount for each step of increased base armor type protection, even before the weapon effect gets involved. (In D&D, we expect a natural 1-point extra difficulty in hitting per step of armor; it's not explicit in Chainmail, but maybe 1/2-point per step in the 2d6 mechanic would be fair.) 

And as a result, you can't just be subtracting from 8, you should be subtracting from 8, 9, 10, 11, or something else, depending on the armor type at the top of the table. Whoops!

Let me be more specific by highlighting a single row and thinking about the story that it tells. Here's the Mace line in Chainmail. It is in fact pretty much just 8's all the way across. The story this tells is, "Screw your armor! Whatever armor you wear, I just ignore it. Maces reduce all worn armor to the same as 'none'". 

Mace 8 8 8 9 8 8 7 8

And here's the row for the Mace in Greyhawk . It is pretty much 0's all the way across. The story this tells is, "The mace is helpless against all armor! Whatever armor you wear, you get the full protective benefit against the mace. Maces have zero capacity to help you punch through heavier armor."

Mace 0 +1 0 0 0 0 0 0

(Remember that you need to flip one of the rows left-to-right to synch up with the reversed ordering of armor types in those two books; I picked this case for convenience in that regard.)

And those are precisely opposite stories, right? Maces reduce all armor to null-value in Chainmail, but give no benefit whatsoever in O/AD&D. And which of those two stories is more in synch with our real-world historical understanding? Chainmail, yea; O/AD&D, nay.

You can repeat this inspection on every row throughout the tables. Wherever Chainmail indicated an advantage over heavy armor, in O/AD&D this turns into no advantage (e.g., the Mace). Where Chainmail indicated no advantage, this turned into a hefty penalty in O/AD&D (e.g., the Spear).

Here's another way of putting this: The factors that go into the Chainmail combat table are weapon type, and armor class. The factors that go into the Dungeons & Dragons attack matrix are character level, and armor class. So when you blindly slam those numbers together, the protection from the base armor class has been double-counted, and every converted weapon in the game looks incrementally worse against heavier armor. 

(Side observations: What's the best weapon vs. plate in the D&D Greyhawk table? The military pick, largely because it wasn't included in Chainmail, so wasn't tainted by the double-counting error, and Gary could pick fresh values that actually made sense for it. Another: Swords & Spells Appendix A has a "% Chance to Hit" by weapon type for normal men table, which largely recreates the format of the Chainmail table -- but the value curves are totally incompatible, because Gary included the D&D modifiers, thereby inheriting the double-counted-armor problem)

Yikes. I'm pretty sure this is the biggest numerical error I've ever seen in the legacy of D&D, and I do think it seems to have escaped everyone's notice for lo these 45 years and counting. 

So I'd say that any 1E players who are still engaged in this gnashing-of-teeth exercise with these tables would be wise to put it to bed, because the whole effect of those tables in O/AD&D was fundamentally broken all along. It doesn't even begin to serve the goal that they're allegedly for. I'm guessing that they were never playtested at Gary's table -- again, he was adamant that he never used them, and was essentially disinterested in the whole project -- but once they got printed and published, everyone took it on faith that they were fit for the purpose. But they very much weren't.

Mea culpa on my own part for not ever noticing this before. I was amazed when this finally dawned on me a short while ago.

To your knowledge, did anyone ever point out this arithmetic-modeling mistake previously? Can you think of any reason to use these tables as-written today? Or would you be amenable to a series of corrected, rationalized tables for this purpose?

Get a spreadsheet for this comparison here.


  1. Never had any interest in using them back in the day, so I don't see a need for a corrected version of the table. I *could* maybe see using 5e style tags to indicate certain properties of weapons, like Mace ignores armor, if they could be reduced to just a few. But I've tried a couple times myself to create something that gave players the incentive to "naturally" choose the historically appropriate weapon that troops used at various periods when different types of armor prevailed, without resounding success. One of the things about D&D is that you find a much greater variety of arms and armor than typical any actual historical period, plus almost all combat the PCs get involved in is free-for-all skirmishing, and then you add completely fantastic elements on top of it. That makes the kind of historical arms races we think we see when we look back really hard to simulate at the low level of detail we (I) am willing to put up with in play.

    1. That makes sense! Certainly a giant table like you get in 1E is really not the way to go to handle it, even in theory.

  2. Indeed, I HAVE noticed this very issue. I ran similar exercises in order to figure out how Chainmail converted to Greyhawk converted to AD&D and arrived at a similar result.

    Here's how I've dealt with it:

    In my OD&D game (that I ran most of last year), I created my own combat tables for weapon vs. armor adjustment. Because of the nature of OD&D (with its limited range of ACs) and because I wanted to use standardized damage (D6s for most weapons, D8s for 2-handers) and because of the (relatively) small list of weapons, I was able to work this out in a satisfactory manner.

    Keep in mind: HTH combat in the medieval/ancient world was not quite so deadly as modern combat! And heavy armor was EXTREMELY effective against these types of weapons. It is HARD to kill a man in plate armor using a sword or (especially) a hand-powered spear!

    [though LESS hard to remove him as a casualty; i.e. as Chainmail would model, leaving out the question of "death"]

    For AD&D, my initial thought was to rewrite all the tables. Then I decided that was insane. Instead, I only apply weapon vs. armor per the rules (i.e. against another opponent wearing specific armor), which has meant the issue comes up fairly infrequently...I don't use it when PCs fight monsters in AC (like hobgoblins) and I do not give the adjustment to monsters at all (versus PCs). I *like* there being something other than damage and weight to distinguish weapons from each other.

    But, yeah, the math problem is glaring, and recently I *have* been thinking of dropping it altogether from the game.

    That being said, I still love speed factor as a method of breaking ties in initiative, and both space requirements and weapon length have solid places with regard to combat logistics (and, for the latter, with respect to who strikes first during a charge).

    1. Thanks for that! You're apparently out ahead of me for noticing that math error.

      Possibly I should have mentioned in the post that my OED house rules have for some time implemented a really simple method: axe-types get +2 vs. chain/plate; club-types get +4 vs. plate.

      I agree with you that this rarely gets used anyway, because combat is so much more frequently with monsters than men.

      And I also agree that D&D combat is deadlier more quickly than real-life combat (I think), and heavy armor is mush surer protection in real life. Likewise, I occasionally think of back-interpolating from Chainmail mass stats (and then realize it's too much of gutting of the system, and not really a trustworthy basis in the first place).

    2. The "Rythlondar" campaign (teh PDFs were available some time ago on the net) had a d20 version of the Chainmail tables, with a "to hit" bonus according to class and level.

  3. This was a very neat analysis! It inspired me to try to figure out what the modifiers "should" be based on the chainmail table, which I posted here: https://iolbs.blogspot.com/2021/03/correcting-weapon-vs-armor-type-from.html

    I don't think, even corrected, that these modifiers make any particular amount of sense. I think the Chainmail table itself probably isn't mechanically very well-considered to begin with, and the situation only gets worse when one takes that and runs with it.

    1. Ha, I was just doing the same thing (and not for the first time) before I popped in here to check comments! Thanks for that link and your work.

      I agree that even a non-glitched-up conversion using CM as a base is on shaky ground. One thing that occurs to me is that the CM table presumably includes extra damage factors for big weapons all in the one roll to-kill (c.f., halberd, 2-handed sword, mounted lance), so in light of D&D's separate hit/damage mechanic, you'd want to separate those factors back out. Among other things.

  4. The real issue with weapons vs. armour in D&D is that combat in D&D is an abstraction. In D&D, two medium-level or lower fighters get one attack every minute versus each other. This does not mean they stand there taking but a single swing every 60 seconds like clockwork. It means that there is a single *effective* attack every minute. The rest of the time they are moving, bobbing & weaving, parrying, blocking, dodging, feinting, and attacking poorly (easily blocked or avoided).

    A mace does not help you attack easier; in fact it gives you a disadvantage. Maces are necessarily shorter than long sword because they are top heavy. This means that the you have to get inside the range of the swordsman *before* you can take a swing. This fact doesn't change depending on the swordsman's armour.

    And then there is the issue that armour coverage is not uniform in D&D. For example, "plate mail" in D&D is a combination of chain and plate. If I were to apply weapon vs. armour modifiers, I have to first check which body part was attacked. Scale (AC 6) is even worse since it varries between scales (better than chain) and no armour at all!

    In other words, D&D combat and armour are an abstraction and using weapon vs. armour modifiers is at an entirely different scale.

    With regards to "D&D combat is deadlier more quickly than real-life combat),” I must respectively disagree. My experience in the SCA has taught me that scoring head shots on low-level fighters is surprisingly easy, having been both a low-level and medium-level fighter myself. In the SCA, a solid head shot from a high-level fighter almost always results in a concussion and a solid dent in the helm -- and that's from using rattan weapons! I hesitate to think what a sharp steel halberd would do! And most helms in the SCA are made of thicker and better quality steel that helms prior to the 15th century. Just one look at the cleaved helms from the Battle of Visby (mid 14th C.) tells us the truth.

    If you want to better model the protection of good armour, use a damage reduction mechanic instead of defense. In other words, have that AC 3 plate mail reduce the damge by 7 points instead of improving AC. In that way, players will be motivated to use two-handed swords and halberds just like real life. Allow critical hits to reflect the successful targeting of weak spots in the armour (eye slots, armpits, groin, etc.).

    1. I think DR is a bad abstraction, at least when combined with D&D combat. It immediately leads to being unable to kill a heavily armored knight with a dagger or arrow no matter where it strikes, which is "fixed" by adding hit locations, called shots and/or criticals or simply pumping up the damage they do, but that makes them supernaturally deadly against unarmored foes, and so on. The basic abstraction of armor making it hard to land a telling blow and random damage to indicate whether it was severe or mostly absorbed/in a non-lethal place IMO just works better with less obvious edge cases than the abstraction that armor means every wound is shallower.

    2. Thanks for the detail on your take on that. Of course, "combat in D&D is an abstraction" by itself isn't very meaningful, because all games are abstractions of some sort; it's really only meaningful in relation to some proposed competing model.

      The truth is, I'm not convinced by the later defenses of the D&D time scale; the man-to-man rules in Chainmail speak concretely of "a return blow", so does the pre-D&D draft document, etc., and I think most of our intuitions are to describe the action at that scale.

      One thing on my mind lately is how the riot-geared officers at the capitol entryway Jan-6 held off an attacking force for several hours of sustained attacks. I'm curious, how would you interpret that in light of your SCA experience?

    3. I think all of this really boils down to what degree of detail are we trying to model in D&D combat. On the one hand we have the 1974 rules with every weapon doing 1d6 damage, no critical hits, 1 attack per round regardless of level (except vs. 0-level) and 1 minute long combat rounds. In contrast we have 3.5 edition rules with variable weapon damage, critical hits by weapon type, feats to improve critical hits, weapon attack & damage, combat manoeuvres, multiple attacks/round at higher levels, and 6-second combat rounds. I think that weapon versus armour type belongs in the 3.5 rules rather than the 1974 rules precisely because 3.5 is a much lower level abstraction. Likewise I think that armour as damage reduction functions just fine in 3.5 but terrible in 1974.

      Personally I am very much torn. One the one hand, I really like having tactical options in combat (both monsters & PCs) and knowing exactly where a given figure is at any given time. On the other hand, with 'theatre of the mind', fast combat resolution + critical hits (and heavy metal music playing) I can (as a GM) facillitate an adrenaline-fueled in-you-face emotional *feel* to combat. The former is interesting and rich but feels like chess whereas the latter is overly simplistic but feels like a Frazetta painting of Conan slaughtering Picts. I really, really want *both* and I suspect that I am not alone.

    4. With regards to the riot-geared officiers, I have seen but not studied the videos so here are some thoughts on riot gear & rioters in general:

      I don't know what level of protection the police wore. The technology available today is simply amazing. There are riot shields capable of stopping .44 magnum pistol shots! However those .44 magnum proof shield are very heavy at 24lbs. Granted, this is only around 10% heavier that what the typical Roman Imperial scutum probably weighed, but from personal experience I can tell you that actively defending with a scutum is very tiring. So given the amount of time the police actively defended I suspect they were lighter and less protective. There are in fact some riot shields as light as 3 pounds!

      Likewise, I don't know what kind of helmets they wore but Super Seer claims that their helmets protect against "high and low velocity projectiles, bottles, rocks, bricks and bullets." (https://superseer.com/Riot-Tactical-Helmets).

      Body armour also ranges from protection from improvised blunt weapons but not knives or guns all the up to pistol proof. As with the shield, the better protection is heavier.

      The capitol riot police had to withstand improvised weapons thrown by amateurs (0-level humans). The helmets, body armour, and shields are designed to protect against precisely these attacks by deflecting and distributing the force across a large surface area. Riot gear is not normally designed to protect against sword-wielding master martial artists, let alone D&D style monsters.

      Therefore I suspect that by the end of the day, the police were fatigued if not exhausted from holding back the crowd. Many of them probably had bruises from bricks or other heavy improvised blunt weapons. In D&D terms I would say that their armour deflected and absorbed most the damage and what wasn't absorbed was turned into non-lethal (subdual) damage. Some of them probably received some lucky shots that I would treat as critical hits which probably did some real physical damage. Also, D&D does not deduct hit points for the actual fighting in combat despite the reality of how exhausting it really is; so I would just rule that once the adrenaline wore off, the police were exhausted. Another way to look at it is that the rioters were 0-level and not proficient in improvised weapons. Under the AD&D rules, non-proficiency gives a -4 to hit. So if the riot gear granted AC 5 or better the rioters would need a natural 20 to hit. Any hit would do 1d4 damage (typical improvised). Alternatively, the AC is 7 but with 4 points of damage reduction so the rioters need a critical hit and roll a 3 or a 4 on the damage die to do any damage whatsoever. Which works out to a 2.5% chance each round of causing damage (natural 20 is 5% × 50% of rolling a 3 or a 4 on a d4 damage die). The actual damage would be 2 or 4 (6 - DR 4 or 8 - DR 4) or an average of 3. So a 1st level riot cop could withstand on average probably 1 or 2 critical hits (1.25) or an average of 80 rounds of active combat. I don't think any of them withstood a barage of 80 hits in a row, perhaps at most 10 hits in a single hour?

    5. Thanks for those insights and analysis!

      Having watched the footage closely (at least the parts of it I saw available), it seems like the lead rioters had come prepared with long poles/stick (which reminded me of the rattan stick you mentioned leaving dents in helmets), and were striking with it constantly, like about every 5 seconds.

      I'm guessing that armored knights vs. levied peasants with spears and pitchforks might be roughly as durable? But of course in its RAW core D&D is more deadly than that in such a situation.

    6. Given how long the police could handle getting wonked on the head, the D&D 1974 rules can only explain it as AC -1 or better (police never got "hit"). Bear in mind that those riot helms may very well have been as good or better than 17th C. pistol proof burgeonets which is treated as magic helms in D&D. As I mentioned previously the helms of the 14th C. or earlier were not sword proof.

      But yes, in general I agree that the 1974 rules do not model hitorical man-to-man combat as well as later iterations do.

    7. Thanks for sharing your insights on that!

    8. Uh... I know that this is an older comment thread, but the Capitol Police had members suffer serious injuries: broken ribs, smashed spinal discs, stabbings, serious concussions, respiratory damage from chemical sprays, trampling, mobility issues, etc.


    9. Well that's a good point, thanks for that. Still the guys in the entryway managed to hold the line the whole time. Obviously huge sympathies to those guys who did get hurt, it sucks so bad.

      (And I routinely get comments on posts from ten years ago, so this still counts as fresh in my book, thank you!)

  5. I forgot to mention, that switching from a 1 point of AC to DR -1 potentially could distort the game. Alternatively, you could give 1 point of DR for every 2 points of the former AC. In other words, plate mail (AC 3) would give DR -3 or -4 instead of -7. Also plate armour has a deflective factor, so splitting the difference by changing plate mail to AC 6/DR -3 may be the best. I've only played with full DR but that was with 3e with tougher monsters.

    1. Interesting option, thanks for that! (Of course Gary started to inject that into the AD&D UA new field/full plate armors.)

    2. I've had a lot of success with DR, using a simplified armor system, roughly on the 2-to-1 scale you mention:

      Leather: DR 1
      Chain: DR 2
      Plate: DR 3

      Shields still makes you harder to hit, but any attack that would have hit you without your shield hits your shield. This matters for things like knockback. The dragon swings his claw with lightning speed and you get your shield up in time -- but the force of the attack also knocks you halfway across the battlefield.

    3. I like that a lot! It nicely equates to Light armour, Medium armour, and Heavy armour. In Red Nails, the "dragon" knocks Conan back 50! So there's your precedent.

  6. Yeah, I was recently experimenting with something similar, and looked at the ad&d table and just ended up ignoring it because it's so obviously garbage and inconsistent with reality. Also as soon as you start adding in dex for ac and monsters with different acs that aren't related to their 'armor' then any table resembling this is thrown out of whack anyway. So I ended up just giving certain weapons bonuses vs chain or plate or all armor.

    1. I ended up doing my own conversion to D&D to hit bonuses/penalties. Any way you look at it, the numbers just don't make a lot of sense. I guess the weirdness of it isn't so noticeable on a 2d6 scale, but looking at it from a d20 standpoint and comparing like to like really brings things into focus. https://42ducktape.blogspot.com/2021/03/weapon-vs-armor-adjustments.html

    2. Nice work and thanks for that link!

      And the other thing (mentioned in a couple comments at this point) with using the Chainmail 2d6 table as a basis is that it wraps hit & damage in a single to-kill roll (e.g., apparent extra benefit to the heavy halberd, 2-handed sword, mounted lance), which we really should disentangle for the D&D system.

    3. Yes, this is something that’s often overlooked. Variable weapon damage, not present in Chainmail or OD&D, is sort of an alternative to the Chainmail kill tables, not a complement to them. There’s no reason D&D needed hit points and weapon damage (even fixed 1d6) on top of the AC system. It could have gone with: roll against AC, a success means you inflict hits equal to your HD (reduced for non-Fighter PCs), every creature can take hits equal to HD before dying. Combine that with a system for attacking multiple targets, for actively defending, etc, and you’ve got a system that approximates Chainmail man to man. Hell, you can then play around with the combat effectiveness of wizard and cleric levels vs certain types of magical beasts, etc.

  7. I never caught this discrepancy, but I had always wondered why two-handed swords had so many bonuses on the weapons vs AC table, and I found my answer the first time I saw the Chainmail table. Didn't think about it much after that though, mostly because I realized that the Chainmail table and the AD&D table are modelling two slightly different things. Chainmail is measuring the chance of "killing" an opponent in a certain armor type while using a particular weapon. In D&D parlance, the chance of doing 1 Hit DIE of damage. AD&D (and any variant using variable weapon damage) is measuring the chance to hit and do a VARIABLE amount of damage, also based on the weapon.

    So I felt that the whole thing was flawed because, using the example of the two-handed sword it get bonusses to Hit and Damage, it really should be just one, the other, or smaller amounts of both. I highly recommend looking at the formula used in 2E AD&D Battlesystem Skirmishes to see how they handled converting Thaco, variable damage, and multiple attacks into a single to hit roles.

    1. Right, totally agreed, and I was just reflecting on that before I came here to check these (insightful) comments. The big weapons in CM all seem to have that same boost (e.g., halberd, 2-handed sword, mounted lance) that would have to get disentangled for D&D separate hit/damage mechanic.

  8. Nice post. I appreciate you noting the fundamental differences between the Chainmail table and the later poorly-considered AD&D table. Although I have not played AD&D in a long time (usually OD&D or B/X), when I did I never used weapons vs. armor or weapon speed.

    1. Thanks for saying that! You're in good company, of course. :-)

  9. Wow, that's an amazing discovery. Given the number of people who purport to play AD&D using weapon vs. armor modifiers, it's shocking that no one has caught that (and publicly reported it) over the decades.

    Yet more evidence of AD&D rules that were not only not used by Gygax, but apparently not playtested or significantly analyzed by anyone else, either.

    1. Indeed! Of course (as expected), when I posted this on the FB 1E AD&D group, there's a proportion of the audience who cherish the text so tightly they're kind of offended that one might claim any error.

    2. There are some real nutballs on there. One guy I ran into clings so tightly to the precious Gygax by-line that he thinks the non-weapon proficiency system in Oriental Adventures is brilliant and the one in 2E is a poor imitation, even though they are almost identical other than differences in the skills available (no charioteering in OA, no calligraphy in AD&D, etc.) ... also totally lost his mind when told Zeb Cook wrote OA and Gygax just put his name on it to sell more copies.

    3. Yeah, that latter one in particular is an incredible sore spot the first time someone runs into it.

      Look further down and you'll see Jon Peterson informing me just now that the weapons-armor matrix may not have been from Gygax, either; mind blown. :-)

  10. I was never really tempted by this table. Initially I couldn't figure out how it can possibly make any sense given DEX adjustments and magic items and the like: I don't think the German translation had a note that "AC" was supposed to be "type of armor" and not "actual AC" number. Later I simply could not be bothered by this kind of complexity anymore. Our games started including some large-ish battles we fought out round-by-round and it could very well happen that an entire day of gaming resolved exactly ONE battle. Who needs complications at that point?

    1. Totally agreed that the whole structure is overly complicated (and clearly we're in good company).

      The "AC = type of armor" is actually a clarification/errata in the later DMG. In the PHB the table literally says "Armor Class" of course; in my DMG it's p. 28 with a dedicated section "Weapon Types, 'To Hit' Adjustment Note" on the issue.

  11. Nice catch there on the math of the conversion. For my part when I wrote up my Majestic Fantasy rules, I decided to incorporate adjustments against specific armor types based on what I read about each type of weapons.

    Partly because it copyrighted and partly because I found the grid form to be cumbersome, I opted instead to write out the modifiers in the description of each weapons and keep it straight forward by focusing on the most relevant weapon characteristic.

    For example

    Mace, small 9d/ea. 3.0/lbs, Damage: 1d4+1
    This weapon is between 18 to 24 inches long and has a ball of metal affixed to the end. It gets +1 to hit versus opponents wearing chainmail or gelatinous creatures like ochre jellies or black puddings. It is usable in the off-hand when dual wielding.

    I have free version of the rules here

    1. Hey, that's a pretty good approach, thanks for sharing that! Definitely an improvement -- and truth is, you've got to start the historical assessment from scratch, so as to disentangle the hit/damage mechanic which is conjoined in CM.

    2. @Robert: Would you consider adding what I like most about the AD&D WvAC table---a club or staff against plate mail is not very effective, i.e. adding some to-hit disadvantages?

  12. Unbelievable that escaped notice all this time. I don't recall ever even trying to use this table as it always seemed unnecessarily complicated.

  13. "On the one hand, they're not listed as variants or optional rules..."

    The 1E DMG at page 28 does in fact suggest that this rule is optional:

    "If you allow weapon type adjustments in your campaign..."

    (Hat tip to Dragonsfoot for refreshing my mind as to where this was located)

    This cryptic reference may be Gygax defaulting to the OD&D assumption that the rule is optional without realizing that it is nowhere else stated as so in AD&D, or it may be that at some point in the writing of AD&D the rule was made non-optional but this reference was overlooked.

    1. Ha! I just happen to have my book open to that very paragraph right this second (thanks to Peter F. above). That's a great catch and interpretation, thanks for that.

  14. Never used the 1e weapon vs AC table, it was just too daunting. That said, have used the 2e version (which is the simpler “weapon type vs AC” variety).

    Just recently learned a simple hack to make that even easier: I pre-calculate the “AC vs slash”, “AC vs pierce”, and “AC vs bash” and write those on the monster stat block (instead of adjusting every to-hit roll on the fly).

    1. Hey, interesting approach! I also use a simpler rule that I apply on the fly (although I'm sure I've overlooked it once or twice, so you may have a leg up on me there).

    2. I knew the 2E version of the table first, and it always bothered me that it had the same problem Dan brings up here where a "good" match-up for the attacker just assumes the armor grants its standard protection (though I wouldn't have been able to articulate that prior to this post). Of the 33 entries in that table, only two (bludgeoning vs. chain mail/bronze plate mail and piercing vs. leather/padded/hide) improve the chance to hit. 17 of them are penalties, and the remaining 14 (including 8 of the 11 entries for bludgeoning) are neutral.

  15. Hey, could you put your OED rules on github? Some of these questions I assume I know the answers to, but might be best made explicit, tersely. I don't know if all of these criticisms are reasonable (nor if page space allows you to address them) but hopefully some of the critiques are productive. Here goes:

    Aside: I noticed that in your treasure determination in the arena program you chose 1d6 x 2 for silver for level 2 instead of 2d6... excellent choice, actually (programming effort vs. variance reduction vs. proportion of total treasure). I don't know if I would have had the strength to do that. :) I don't know Java, per se but I've found it helpful to look at the code. Also, I don't know if this is great in the context of what is to come next, but I very much appreciate what you've done in these rules and on your blog.

    General: *Link* to your blog posts, please, vs. just refering to them by date.

    Players rules (1.0.6): What does light armor mean? Do fighters get +1 attack bonus at level 1? Does attack bonus mean to hit? Does the "1" or "2" results include con bonus/penalty or not? What does a potion of mithridate do?

    "For most combat actions, roll d20 + level +" should be "+ level bonus (based on class) + enemy AC" ? Or something to that effect? As written, it sounds like wizards get +3/2 per level, and fighters get +2 per level, instead of +1/2 and +1. If you don't want to get into it, in the interest of space, just say that you use target 20 and give it a 6.

    "The multi-classed character uses the best entry for armor, weapons,
    attacks, and saves." --> "for everything" to save space, if required. At the very least, remove the "and saves." part.

    "Fighter/wizards may cast spells in leather or chain (melee weapon in one
    hand, no shield)." Can they use torches and/or magic orbs in the other hand?

    "Thief skills are restricted to leather armor only." This is my precise reading of the 2e rules, i.e. that you can wear plate as a thief but you can not pick a lock while wearing said plate... but I find that most DMs read it as "thieves may not use any non-leather/non-light armor". Feel like clarifying that (in your rules)?

    "(3 tries allowed per target)." Does this include hide in shadows? Does a thief have 3 chances to get caught during pickpocketing?

    "Thieves add +1 to hear noise and climb wall checks, and score double damage from a silent attack from behind. These bonuses increment every 4 levels." Hmm... ok, so x2, x4, x8? :) If you just say x2 then incrementing means x3, but if you say doubling, then I have no idea what incrementing means (note: I haven't done well in high school math... ok, so the funny thing is that I put this post into a character counter on the internet, since I think blogspot only allows 4096 chars, and this was, according to Dale-Chall, 11th-12th grade writing, eerie...). But if you write something like "Thieves add +1 to hear noises and climb wall checks, and do an additional 1x damage from a silent attack from behind..." I don't know, maybe this can't be fixed. I feel like, as written, it could easily become best for a high int person to find the traps and then a thief to remove them, but I don't know if that's intended...

  16. If you intend to let people reroll 3x upon level-up on learning spells or picking locks, etc. then it may be good to indicate such. If there isn't uniformity between those elements it may be useful to clarify that. Can you attempt to research a spell 3x?

    "Attack rolls are –1 per 10 feet range." Change this to "per 2""?

    There doesn't seem to be a reason to use longbows in your rules, other than, perhaps, prevalence of long bows vs short bows in terms of magic items, in basically every game that I've ever played in. Further with crossbows firing once per two rounds, maybe the value comes from encumbrance of the quarrels (while holding swords?), it doesn't seem good to use them. You mentioned bow < crossbow < rifle, in terms of training a lower class urbanite to be combat effective, but there's no mechanic for that in-game, afaik.

    "Halflings can only use light or medium (two-handed) melee weapons, slings, shortbows, and light crossbows. Thieves are skilled in light melee weapons, normal swords, slings, and light crossbows." --> "Thieves can use light melee weapons, mundane swords, slings, and light crossbows. Halflings can use light weapons normally, medium weaopns with two hands, slings, ... " or state one-handed swords vs. normal swords, if that is your intention.

    (Officially drunk now. This may trend towards the obnoxious, but I hope just a little bit will be reasonable, or at least reasoned.)

    "play goes around the table." Great idea. Remove. I feel like this is a trapezikentrikos idea that doesn't need to be here, but ... I guess. Whatever man. I think you mentioned bonus XP for seating arrangements at a trapezi, and that's reasonable, but still.

    "Attacks: Attacks are made by rolling d20 + attack bonus + AC of target and scoring 20 or more. Monsters use their hit dice for attack bonus." Please forgive me, but I have no idea why this is here; isn't this #6 again? If this _is_ here for a reason, then why is it there? Let's call this 20.5 and just go straight from 5 to 7. Or, what the hell, call it 6 again. That'd probably be better (though this blurb seems better than the 6 blurb).

    "resist magic +4" --> save vs spells? Maybe I'm wrong about this. Anyhow, please clarify.

    Fighter Feats. Sad but true, never had a character achieve level 4. Skipped. Insert jealousy here.

    "Elves automatically start with the wizard class," --> Honestly, this seems way way way different than I would expect from a lore perspective (everyone should start fighter, right?) BUT on the other hand, this is admittedly better for the game than starting with fighter, as M/T > M/F, in terms of uniqueness. I would, however, say that you should make it explicit that people can take 3 classes instead of just 2, and... although this is highly stupid, and probably worthy of ignoring, since race is a class, state that you can't multiclass into a race? Meh, ignore that. But ... well, race as class, right? :) Just nevermind this, sorry man. :) To be fair, if you allow, as Gygax did, learning foreign alignment language, then multiclassing into a race doesn't seem crazy. :)

    Adding Classes: ... Experience: ... Benefits: ... if you require the space, you could combine some of these things into one paragraph. As written, though, if you have the space, it's lovely and well laid out.

    "...armor, weapons, attacks, and saves." What does this not include? Encumbrance? Removing traps? Can non-thieves remove traps? Anyways... you can probably do something with this that makes it more terse.

  17. "the best entry... Hit points are rolled for each class..." I actually like this, as it specifies 1d4 vs 1d6 meaning the best result, vs. the best choice between them.

    I made it to page 4/8 here and I didn't review the rest yet but I know that I've seen at least one typo or spelling error that I haven't mentioned here yet (if you care, I will find it), which has existed for a long time in this document.

    (While not drunk..,)

    "Warning (e.g., light, noise, magic detection) foils surprise. " --> pretty subjective, but "Warnings (e.g.... ...) foil".

    Judges' rules: "minus encumbrance level" Do you mean encumbrance number/value/amount? What does "level" mean here, like 1 = 1 encumbrance, or 1 = 12", 2 = 9", etc.?

    Rest of judges rules were not reviewed in this post.

    (Back to drunk posting again.)

    Anyways, again, thanks for the awesome rules. I really, really appreciate it. It's amazing, and I appreciate the code on github too, even though it's Java... and I like your math posts. I just like you in general. I'm an armchair GM for the most part, but I figured I'd interject and if you take like one correction out of this that'll be fine.

  18. Feel free to tell me not to comment in the future and I'll respect that. It will save me some time for other pursuits. :) Again, thanks and take care.

    1. I'm delighted that you've taken the time to read my stuff in such detail, thank you kindly! For time/space purposes I may need to offload answers to some separate FAQ posts, perhaps (copy-pasting your questions into my tasklist right now...)

      Do you think there's advantage to putting the OED house rules on Github instead of just the main website at OEDGames.com? Honestly, there's was a prior point where I hosted more non-code documents there (e.g., presentations), but later it seemed to not fit well with the idiom, so I removed them.

    2. There's some typo/spelling error in there somewhere that I can't find now but which would be more easily fixed via github than blog comments. That's the sole advantage as far as I can tell, and you can submit "issues" there, though to be honest I am not great with github/git (it likely end up being a pull request made via in-browser editing).

      As an aside, when I wanted to look up how you did treasure I found your code to be really useful (I use python so it was neat just to see some java). I noticed you shortened the level 2 silver to 1d6x2 (vs 2d6x1) and I thought that was damned brilliant in terms of coding efficiency vs. EV vs. the very small amount of variance exposure. An excellent, rational, short-cut. :)

    3. Thanks for saying that! Yeah, as a math guy the asymmetry of the Vol-3 book treasure value at level 2-3 just hurt me to look at it (I likewise have it penciled in in my paper copy). So glad having the code in public was helpful!

    4. Ok, so in the above I haven't been commenting on the rules themselves, just the writing, but I'll take a crack at the first 4 pages from a rules perspective now. Some of this is criticism and some of it is just questions about how to run the game. There's like to be some repetition here with the above so I apologize for that. Some of this is just stuff that I don't know how to run, period.

      Polearms appear to be dominated by pikes, including that they cost more. Maybe give them a d10 for damage, or give pikes extra encumbrance? You seem to want something thats sims well and scales well, with an emphasis on something that is fun to play as an RPG, whereas I want something more like a board game with RPG dressing. I can appreciate that this might cause complications at the mass scale.

      I don't know if I said this already, but I don't know if starting hit die "results" of 1 or 2 includes con modifier. It could potentially put you in an infinite loop with a mage with 3 con, for instance. :)

      A potion of mithridate should be briefly explained here, maybe even by changing the name, as it would be more terse.

      I'm concerned about clubs and especially axes really leveling out the AC of things. But I haven't looked at it carefully so this is just intuition.

      I don't think anyone's effectively addressed where a door is at the 1" = 10' scale (standard scale for drawing on graph paper?) and how you fight in/around a door. I think I'm of the mind to just ignore the doorway and treat it as an open hallway, for the purposes of movement and combat, even though it's totally ridiculous to do so.

      With ranged attacks at -1 per 10', is there ever a good time to use them, vs. just advancing to attack? If you are stationary and the party moves forward, you can still fire into that melee "safely"?

      How do you run the acid breath on a giant bombardier beetle? I don't understand the entry in the monster DB. Does their AOE increase their EHD? Also I feel like the Giant tick is way underrated, as it has such low AC and high bonus to attack. EHD 2 seems too low, but maybe I'm wrong, based on the sims.

      When a player chooses a stat for level 1, should they always choose dex?

      Can you hotlink your footnotes in the PDF easily? Page-up/page-down is killing me.

      It seems like you are separating race and class, but it's not clear. Can a halfling multiclass?

      The idea of gradual slopes that move an entire dungeon level in under 100 (or even more than that) are nuts. I mean you can just look backwards or forwards and see it. So let's call this a mythic underworld and accept that it works that way, but mechanically, how do they note slanting passages (automatically)? What's up with these "slope so gradual even a dwarf won't notice" designs?

      "Warning (e.g., light, noise, magic detection) foils surprise." You might be able to, if you want, squeeze listening into this space.

      Thief skills restricted to leather armor... can they wear plate otherwise? What if they're not wearing armor?

      Can an elf multiclass F/M/T?

      Berserking, by turns, do you mean rounds?

      Great fortitude: I've been giving death saves vs. poison. Is that accurate? Given that there is no poison save in your chart, maybe poison is a breath save and this only gives you a bonus vs poison saves, while using the breath save? In any case, consider putting something in parentheses to indicate the base save for poison.

      Fighter feat: Toughness, retroactively?

      Regarding two-weapon fighting, and its reference to Chainmail, since there are so many versions of chainmail, which one are you talking about?

      Again, that takes me to page 4. I love the feats, to be honest. Given that you get one at level 4, it gives you just enough Halfling to get rapid shot. :)

      This is considerably more poorly written than the above, so, sorry about that.

    5. I hadn't considered halflings with polearms, though, which may be a reason to not increase the damage on them.

    6. There doesn't seem to be a disadvantage to using a shortbow (unless you come across a magic longbow).

    7. Ok, so regarding the advantage of swords (use immediately). I'm curious if the 1d4 dagger is enough of an edge there (I think it is) vs a short sword, but I also don't want to see people throwing short swords (or maces, tbh). :)

      Do you follow the rule that players that are surprised can drop what's in their hands (this would improve the value of swords)? Does a torch go out if it's dropped (as per gust of wind)? It seems like a lot of rolling to do the drop chance and then determine what hand drops its contents, and it seems unfairly balanced against the players.

    8. (vs. a 1d6 short sword)
      *nor maces.

    9. Good questions there, I've added them to my FAQ-todo at some point task. Just on the very last item: I actually don't use the surprise-drop-in-hands rule, largely for the bunch-of-rolls issue you mention. Although PCs can drop an item at will and pull out a sword, which happens quite a bit.

  19. The greatest annoyance to me with the weapon vs Armor chart is it didn't do what it said it did. It doesn't reflect any sort of "realism" of weapon characteristics vs armor characteristics, By example in AD&D any goven AC score is not just one armor type. It's multiple types of armor and as the AC score is now a score and not a fixed index as it was back in chainmail what it was tryign to do simply wasn't happening. By example AC 4 is multiple types of armor and DEX scores all jumbled together. To properly get the right modifier you need AC without shield, AC without Dex modifier, AC without magical adjustment and you still have the issue of picky that AC 4 can still be multiple different types of armor. I like the concept but it just doesn't work given what AC represents in AD&D.

    1. Now admittedly on that issue there was a clarification in the DMG p. 28 (it's come up above a few times at this point) that the weapon-vs-AC table was meant to represent the "specific *type* of armor" only, before any modifiers (such as Dex). But the with/without shield case you're spot-on about.

  20. Combat rules in any RPG have been this tug of war between people who want the ultimate simulation, people who want lots of tactical options and doodads, and people who just want a simple and roughly reasonable combat system with room for a few tactical decisions so they can focus on other facets of roleplaying. AD&D as written capitulated to the simulation nerds by incorporating and imitating various realism bolt-ons from the supplements, but only partway and in a way that ultimately satisfied nobody.

  21. I've liked the idea of weapon vs armor. But I always thought the numbers in that chart were way off and stupid (never thought systematic error, just caused by the various myths and misunderstandings about weapons and armor in 70's/80's, some of which sadly persist today).

    So, I always started rolling my own. Then I thought, "no one likes this system, they're gonna be even less interested in my house ruled version", so gave it up. Later I quit trying to make D&D a simulation of medieval combat and just enjoy it for the fun game that it is.

    1. Definitely agree with your sense in that first paragraph there. Actually I guess I agree with all of it. :-)

  22. I love this post. Probably the best thing I've seen on the internet this year. I appreciate your inclusion of the original tables and the spreadsheets you used to do your analysis.

    Nevertheless, I disagree with your conclusion. You say that the mace in Greyhawk "is helpless against all armor! Whatever armor you wear, you get the full protective benefit against the mace. Maces have zero capacity to help you punch through heavier armor."

    And then because of the similarity between the Greyhawk and AD&D tables, you seem to be saying the same think about AD&D.

    But that's not how the AD&D table works out, after you apply these bonsues/penalties to the to-hit tables on pages 74-75 of the DMG:

    Both maces (footman’s & horseman’s) get a +1 against AC 2. Meanwhile, the short sword and broad sword are both -3 against the same AC and the long sword is -2. A 2nd-level fighter needs a 16 to hit AC 2 with a mace. With a long sword he needs a 19 and with a broad sword or short sword he needs a 20. The mace is doing 3.6 times as much damage as the broad sword against plate and shield.

    The same is true for AC 10. The same fighter needs a 10 to hit an unarmored target with a footman's mace, but he only needs a 7 to hit with a sword (broad, long or short). Now the broad sword is doing is doing 41% more damage than the mace.

    Gygax wanted to convey that maces were better than swords against heavy armor, and worse against no armor, and the PHB table does that.

    1. Thanks for this, and the similar conversation on Facebook. I'll report what I put over there because I think many people on FB reached for the same counterargument:

      I think that the contrast between two weapons is a red herring here. The conversion error translated all weapons by the same amount, penalizing everything vs. heavy armor, so those relations stay the same and aren't interesting.

      The key thing is that in Chainmail a mace negates heavy armor, and in D&D it doesn't. If Sir K. knows he's going up against a mace-wielder, in D&D he might as well put on plate b/c he gets basically the full protective benefit. Whereas in Chainmail he might as well wear nothing and be fully mobile, with the same defense as plate. That's a huge difference of intent (and it's repeated for all the weapons that should be penetrating vs. heavy armor).

      Thanks again for making me think and express that clearly!

    2. I think you're making my point for me. The idea that Sir K. would go into battle naked because his opponent is using a mace seems silly to me.

    3. It seems completely intuitive to me. Each step up in armor is considerably more heavy, hot, tiring, and expensive than the previous, so there's little point in wearing it if you can similar protection without it. Simply using a shield and a good/versatile weapon was very much a viable option for a long time, and I'm not sure off-hand if heavy armor ever really became common outside of ceremonial use in regions without the iron deposits of Europe, even after trade routes were established.

    4. And it being silly or not to you is another distraction. The point is: It was a big change. Gary explicitly wrote, and the numbers also demonstrate, that he did some simple conversion trying to recreate the effects of all the weapons from Chainmail. The fact that it reverses all of the effects of all the weapons against heavy armor simultaneously is very strong evidence of an accident on his part.

  23. From Reddit without corroboration. Interesting if true.

    "Dave Arneson taught one of of my classes. He told us an anecdote about the D20. Basically it boils down to this:

    Dave was a big wargame player, and often tinkered with his own rules sets. One day (in his teenage years) his family went on a vacation to Europe. He went into a games shop while he was there and they were selling loose D20s. He was excited about what he could do with that sort of die, so he purchased a couple. Later on, when Gary showed him D&D, Dave started using the D20 with that game. (Gary, with his Chainmail rules, had been using D6s still.)

    When they (then TSR) began plans to publish a new edition of D&D, they started looking around for a US supplier for D20s to include in the boxed set. (I believe this would be the Holmes Basic boxed set.) The best deal they could find was with a "learning supplies" company that produced a whole pre-packaged set of polyhedral shapes. They couldn't order single D20s--they had to buy the whole package. They hated the idea of having to discard all of the non-D20 polyhedrals, but came up with the idea of using them as variable sets of damage. Thus, the whole range of polyhedral dice were included.

    So, to answer your question, the very original D&D manuals came with two sets of rules: one that uses D20s (for those who could get their hands on it,) and one that used the Chainmail wargame rules with just D6s (for those who couldn't get their hands on D20s.) The next edition of the basic set added the other polyhedral dice."

    1. That's an interesting anecdote, but it seems to be exactly reversed from what people who played with Gary in the earliest days all swear to (e.g., on ODD74 boards). Those players are adamant that they never used the Chainmail d6 mechanics ever in playing D&D, that it was inserted into the rules as a marketing ploy to sell more Chainmail.

      And everyone seems to agree that there was one meeting on a particular date where Arneson and Megarry showed Gary about playing a dungeon-crawling game, not the other way around.

      And all editions of D&D (including OD&D and even the pre-D&D draft) definitely feature all the polyherdal dice. So that Reddit account is very untrustworthy.

      If you don't have OD&D yourself, then I'd highly recommend you pick up a digital copy! It clarified many issues for me -- and that excited revelation was the thing that motivated the start of this blog.

  24. Does anyone know why Gygax and Arneson switched from 2d6 to a d20? I don't remember it mentioned in historical accounts (e.g. Playing At The World, Empire of the Imagination, Of Dice and Men, etc.)

    If this history isn't already established, I feel like you may have stumbled onto the reason: they stared at the Chainmail table and realized they didn't have enough granularity so they moved up to a d20.

    This is certainly what I would have done (and what I have done). When I'm not getting the right amount of granularity from a mechanic, I switch dice.

    In Chainmail, the average to-hit number for AC 9 is 6.92 — the equivalent of needing a 9 or higher to hit on a d20. For AC 2 the average is 9.54 — roughly equivalent to needing a 16 or more to hit on a d20.

    So, converting to a d20 gives you the following, before any adjustments for weapon type:

    2 AC = 16 to-hit
    9 AC = 9 to-hit

    16-9 = 7
    9-2 = 7!

    This means that if you use a d20, each step down in Armor Class makes you exactly one point harder to hit. In fact, you could use the following formula:

    Number needed to-hit on a d20 = 18 - AC.

    And then when you need modifiers by weapon type, you go back to the Chainmail table. With a d20, the effect of these weapon type differences will be reduced by 63%. But that's a good thing, because people were complaining that a mace is more effective against a knight in plate mail than against a naked man.


    1. The thesis in Playing at the World is that (a) a move away from a bell-shaped distribution makes it easier to assess the effect of modifiers, and (b) lots of wargamers were working from real-life statistical data charts (e.g., Korns in Modern War in Miniature), and doing all kinds of gymnastics to map that to d6's, whereas d20's in 5% increments are a much more straightforward map.

    2. One thing I might add is that I don't believe the Greyhawk weapons vs. armor charts (p12-13 of 1st GH) were actually created by Gygax. My understanding is that Jeff Key contributed those.

    3. Oh, holy smoke, really!? Wow, thanks for popping in and sharing that.

  25. The 2E DMG has a smaller table (3-columns) that has damage type vs. armor type (e.g. Piercing, Slashing & bludgeoning vs. Ring Mail, Chainmail, Banded, etc.).


    1. Right, many people find that both more reasonable and easier to use. Of course, I use an even more cut-down version of the idea in my OED house rules.

  26. I have designed but never tested a combat system which is basically a flat target number to hit, with weapons types (slashing, crushing, piercing) doing different damage to different parts of the body and armour absorbing damage (f.i. chain absorbs 2 piercing and 4 slashing). The idea is you'd either roll the damage die on the table, wherever it falls meaning where you've hit the opponent, then quickly substract the defence of the armour and reading the result (which is easy to understand, like "broken rib, -2 to rolls" or "dead in 2 rounds"). If you want to aim for a given part of the body you roll with penalty or something. Weapons have mass (more mass, more damage but less speed) in addition to shape (reach and damage type) which should be easy to grok too.

    I am really looking forward to testing it...

  27. Delta, thank you for posting your observations. Your pieces on Chainmail over the years continue to draw me in and often illicit response.

    I have to admit I am a bit torn by both your argument and that of Clay Dreslough.

    Seems to me this IS a question of translation, as much of the MTM as that of Chainmail's MTM to Supplement I. I have often wondered why a dagger and a two-handed sword performed equally well against an unarmored man, while a battle axe or a mace performed poorly. Clearly the table takes into consideration, as you rightly observe, more than just the effectiveness of heavy armor. One might counter considerations other than heavy armor are already assumed by Chaimail’s ‘weapon class (1-12)’ mechanic. Whatever the answer, all that was subsumed was more likely ‘eye-balled’ or geometrically rather than arithmetically deduced, there remains an ambiguity about what is assumed in the abstraction of the MTM. All of that perhaps, points to a weakness of the inclusive numbers in the MTM chart or perhaps a strength.

    Further, I may not fully comprehend your point, but does not Clay Dreslough's contrast of swords and maces from Supplement I through AD&D 1e provide a cogent example of weapon factors derived from more than just a relationship to heavy armor? As I read it, maces lack an advantage against unarmored opponents due to their unwieldiness, whereas swords, here, excel in the hands of a proficient regular or better.

    1. Maces lack an advantage against unarmored people because there's very little you can do to reduce blunt force trauma with armor. Even with curved/sloped surfaces to reduce the normal component of the impact force/momentum, you're only looking at less than 10% reduction (e.g. an impact vector that's 20 degrees off from being normal to the contact surface would only "lose" about 7% of its normal force). Points and edges are far more sensitive to that impact angle.

      In other words, if you're going to adjust for weapon type vs. armor type, doing it sensibly would end up with armor having little impact on "heavy thing on a moment arm"-style weapons (mace, hammer, flail, etc.).

      The Chainmail chart reflects that. The AD&D chart does not, because it should be giving those weapons a bonus to-hit to offset at least most of the armor's AC, whereas the table shows most of those as being about neutral across the board (which would reflect being equally effective against all armor types, a very different conclusion from ignoring most of any armor's effectiveness).

    2. Yeah, Ash said it with more detail than I could muster there.

    3. I do wonder why people believe this. I've fought people with maces and I can tell you that a mace that puts a dent in plate armour and knocks the wind out of you will puncture lungs with bits of your ribs if you take the armour out of the picture.

      Armour can be very very effective at reducing blunt-force trauma. Even blows to the head, which is still vulnerable unless you have good neck bracing are significantly mitigated by good armour.

      If anyone wants to prove me wrong, I'm happy to be the one with the mace.

  28. Ash Adler wrote:
    "because it should be giving those weapons a bonus to-hit to offset at least most of the armor's AC...

    Huh, I expect it is as Delta describes, the mace entry in Supplement I and the PH turn out to be nearly the inverse of the entry in Chainmail.

    Strange, the error is so uniform it's as if the editor conflated the weapon with the armor.

    Redress would mean flipping the values. I like the redress Wazbar (above) offered in his blog post, though I disagree with him about the mace, again there is more than toughness of armor assumed in the MTM abstraction.

    1. I'd guess that Dan's theory in the OP is probably right: it was an oversight by whoever made the Sup. I table to forget to-hit in Chainmail combined weapon and armor into a single target number while to-hit in D&D has a separate step for applying the armor's protection to the target number.

  29. One other, perhaps minor, oversight in the conversion from CM to D&D, was that in CM a successful hit = a kill, whereas in D&D the hit simply inflicts damage (which may or may not kill). I did always consider it a great flaw in CM that a hit was an outright kill - wounding should have been possible. The Boot Hill rule of "light wound", "serious wound", and "mortal wound" ought to have been used, and would have worked well with the D&D Cleric spells of "cure light" and "cure serious" wounds, perhaps even obviating the need for "hit points" altogether.

    1. Good point! My understanding is that was a point of contention for almost the very first RPG combat Arneson ran, causing them to house-rule in the idea of hit points then.

  30. I don't think it was an oversight, rather more so an affect of the game being a simulation. I expect most were hobbled or succumbed to injuries on the medieval battlefield. Nonetheless, I share your feeling on this, we altered the kill dice to a wound/kill dice and were able to keep the same probability distribution.

    1. Yeah, it did occur to me that "Kill" could be expanded to mean "Killed or Incapacitated" but it is possible to be wounded in such manner that one is capable of fighting albeit in a diminished manner. Having said that, though, in the adrenaline rush of a real fight one would probably not even notice "light" wounds (until AFTER the fight...) and a Serious wound would probably incapacitate, so the abstraction in the Chainmail rules is in fact acceptable. One could expand it, of course, but for a quick and dirty sort of thing it is acceptable.

      But if one wanted a more detailed treatment I would suggest that if you rolled exactly the target number needed to hit, it counts as a light wound. Exceeding would be a Serious wound, and exceeding by "X" amount (say 3 or more?) would be a Mortal wound (alternatively I might assign a special target value for each weapon/armor target to score a mortal wound, reflecting that a sword slash is more likely to inflict a lethal wound than, say, a simple club.)

      Light wounds might require a morale check to avoid running, though otherwise not impact a figure (unless one wants to track the accumulation of light wounds, based on some sort of "level" system and getting so many causes minor penalties to hit and movement). A serious wound probably incapacitates outright, but if it doesn't expect significant movement, etc. penalties. Mortal wound is self explanatory... :-)

      Could be an interesting system - I'll have to give it some thought.

    2. Angantyr wrote:
      'in the adrenaline rush of a real fight one would probably not even notice "light" wounds (until AFTER the fight...) and a Serious wound would probably incapacitate, so the abstractionin the Chainmail rules is in fact acceptable.'

      Like you said this is reminiscent of the Boot Hill rules which are indeed more like a simulation, e.g. a serious wound is a more likely outcome and without treatment one will succumb to the injury.

      'Light wounds require a morale check...serious wound...mortal'
      Yeah, that's pretty much the shape of it, as I see it, though morale checks on light wounds should not be required of figures with a morale of 8 or more. I would only track multiple wounds when a serious wound is incurred.

    3. Those are nifty ideas there for the Chainmail system, I like that stuff.

  31. Truly fascinating! I never used the weapon vs. AC tables in the old days because they always seemed too cumbersome to be worth the bother, but that also meant that I never cared to take a hard look at them until quite recently, namely when I ran AD&D for the first time in several years this past January.

    I was going to use the tables because I wanted to play 1st edition "by the book," only the modifiers were so patently absurd that I wound up electing not to use them after all, because they once again didn't seem worth the effort. But I never suspected systemic error. It just looked at a glance like the typically inaccurate understanding of medieval weaponry I've come to expect from games written in the 70s and 80s.


    In the recent past, I have used a simple house rule in my Basic D&D games, where common one-handed melee weapons were each given their special situational advantages: namely, spears could be thrown, swords and axes got +1 to hit against "soft" ACs (10–6), and maces, hammers, picks, and flails got +1 to hit against "hard" ACs (5–1). And even this simple rule wound up getting dropped eventually, because even though I wasn't even using Dex adjustments to AC, it was just too much trouble to disentangle shields and magic adjustments from the "base AC" that triggered which weapon got the bonus to hit.

    Certainly, the rule could have been formulated in terms of the in-fiction armor type rather than a number ("slashing weapons are +1 to hit the unarmored and those in gambesons, leathers/hides, or lamellar; bludgeons are +1 to hit foes in brigandine, mail, lorica, or plate; pick whichever 'armor' type makes sense for monsters"). But for a paltry +1 bonus? I was as inclined to forget that in the heat of play as I was the +1 bonus to missile attacks at close range that also didn't long survive contact with my table.

    1. Thanks for the compliment! Yeah, I agree with basically everything you wrote there. I have a dogma now that I really don't want to deal with situational modifiers of only +/-1.

      My OED house rules have used this for a while now: "Swords can be drawn and used in a single round. Spears can be used to attack from a second rank. Axes get +2 to hit targets in medium or heavy armor. Clubs get +4 to hit heavy armor."

      Combined with the varying damage amounts from Sup-I (I only use the first column), every weapon has a unique profile.

      Admittedly, it's halfway rare for PCs to fight other humanoids in heavy armor, so these modifiers don't come up a lot, and sometimes I'm on the cusp of forgetting them when things get heated.

  32. Good post. I worked up a chart that preserves the probability to hit from Chainmail though I don't think it's a particularly good idea, in the absence of the weaknesses of heavier weapons in Chainmail (like letting opponents using lighter weapons parry).

    1. Thanks for that! I also went down the same path recently, and then gave it up for similar reasons. (Actually that was one of the things I did first before realizing the specific issue that generated this blog post.)

  33. I'd like to say I see the problem, but I don't. Perhaps it's me or just the way it is worded:

    “Even if your "base to hit" in Chainmail was 8, it shouldn't be 8 all the way down the line. It should be naturally adjusted by some amount for each step of increased armor protection, even before the weapon effect gets involved”.

    I am inferring the ‘it’ in ‘it should be naturally adjusted’ refers to the base score derived from the MTM. If so, that number, as you acknowledge, is inclusive of all the armor types on the MTM table including no armor.

    One could choose to arrive at the average or base score for an individual armor type and using your formula, subtracting each number in the column from that base score ‘to hit’, after determining the average.

    Whether of target numbers is derived from the aggregate of all armor columns, or, as in the case above, just one column, the number arrived at is inclusive of the armor and the weapon. Therefore the resulting modifier for Supplement I must be inclusive as well.

    Therefore, how can the weapon factors table fail to assume armor if the table from which your formula is derived does exactly that?

    The problem of the mace seems like a valid and yet altogether different problem.

    Perhaps, you might be more explicit as to your conclusion that the implied protective value of the armor in the MTM is not recognized in the tables Supplement I?

    1. "Whether of target numbers is derived from the aggregate of all armor columns, or, as in the case above, just one column, the number arrived at is inclusive of the armor and the weapon."

      But those two methods would generate very different values. The average-per-column method would have been better (but still biased due to inclusion of the weapons that are serving to negate the armor).

      "The problem of the mace seems like a valid and yet altogether different problem."

      But the mace is just one easy-to-read example that I've highlighted for people having trouble reading the statistical table output. It's all through the results for every weapon. The next-most-evident cases I'd point to would be the battle axe, flail, and pike.

    2. The claim, as stated, is a base score of '8' is erroneous, as there should be a base score for each armor type subsequently modified by a weapon. As you observed, the value of individual armor types are made plain in the alternate combat system,each armor type adjusting the base score 'to hit' by 1 pip. The weapon type is also pulled from the abstraction of the MTM and made visible as a weapon factor modifying the 'to hit' dice. That is, in Chainmail the final target number is not a matter of arithmetic as it is in D&D, rather the target number is in a sense, a priori, the basis for the 'disentangled' arms and armor of Supplement I. The numerical value given to the armor and it's relationship to a melee/missile weapon are assumed in one number. Further, as you attest with your formula, the numerical value on the MTM is likely an average derived all armor types (including no armor) vs. weapons.

      We also appear to agree on the difference between the two methods and a 'column by column' vs an average of all weapons and armor types to achieve the base score.

      "...biased due to inclusion of the weapons that are serving to negate the armor".
      Here, I don't follow the argument. Is not the base value of the armor buried somewhere in the average of the target numbers for each column and arguably represented in the average of all target numbers in the MTM table? Would it not be an absurdity if it were not so?

      Again, the mace, the battle axe, the flail, the pike etc suffer from a problem you soundly describe. The conversion method you have surmised in Supplement I fails to recognize the implied value of the armor type in Chainmail; c.f weapons, like the mace, that reduce armor to little or no value in the MTM are nullified while those with little or no advantage in Chainmail are given substantial penalties (e.g. the spear) in Oe and AD&D

      Agreed, Chainmail is the more likely simulation of our world historical understanding. Though it would be the subject of another thread, the realism attributed to some of the target numbers in the MTM is debatable.

      Conceptually, these problems appear to stand apart from one another, as one could exist without the other. Despite the obscurity of a base target number for each armor type in the MTM, I am a bit confused as to how else these numbers would be derived inherently, as it seems those values are contingent with the weapons and perhaps the other armor types.

      Edits: my apologies for the multiple posts, it is annoying that a post cannot be edited without deletions.

  34. I *knew* there was something deeply wrong with the weapon v AC mods, but never saw exactly where the math went wrong. Well-spotted! My current super-simple mods are: swords +2 vs unarmored (including most monsters), axes +2 vs leather, spears +2 vs chain, and bludgeons +2 vs plate. AFAICT they're basically realistic and the symmetry makes them easy to remember.

    1. I like that, very practical! I have kicked around giving axes a bonus against shields, but the fact that shields are only +1 makes it a little tricky.

    2. Thanks for saying that! I agree, and my own OED house rules are similar in tone: axes get +2 vs. chain/plate, club-types get +4 vs. plate. The advantage I give to swords is they can be drawn and used in a round, while other types need a round to get loose.

    3. In my BECMI game I'm planning to give swords and barbed-head arrows + 1 against anything less than scale mail, while bludgeons get - 1 in that range. Then they flip modifiers for anything above scale mail. Meanwhile axes/crossbow bolts sit right in the middle with no modifiers. Finally, spears get no mod until heavier armor, when they are like swords in melee (-1), but like bludgeons when charging or set against charge (+1). The nice thing about having scale mail as the inflection point (where no weapon has a mod) is that most BECMI humanoid monsters have AC 6, implying scale mail, so most of these mods are more the exception than the rule.

      PS: This blog is a major inspiration for my own house rules; thank you for doing it!

    4. Wow, thanks so greatly for saying that! Totally made my day. I really like your take on the weapon adjustments, I'd be happy to play with that. :-)

  35. Do you have a FAQ up for ODD yet? I haven't been able to find one.

  36. Ah, I see you mentioned making faq posts (new) above, if I'm not mistaken, so if they haven't been done already then I guess just delete my comment here and above and I'll see them when they get there. Thanks. :)

    1. Right, that's still a pending work. Thanks for asking!

  37. This is an AMAZING find! Thank you for sharing it, with spreadsheet and all!

    1. Thanks for saying that! I kind of gasped when it dawned on me myself, in fact. :-D

  38. Excellent post! Since I run the LBBs + Chainmail with all weapons doing d6 damage, the use of weapon class for first strike/parrying from Chainmail has been useful to vary things but the different capabilities against armor have been difficult to capture. I don't want to disincentivize using longer/slower/heavier weapons by not capturing the things they are good at!

    The Greyhawk/AD&D tables are too much (and borked as you point out) and as much as I love Chainmail, porting the 2d6 targets to d20 modifiers gets crazy or unintuitive results (like a 2-hnd. sword getting a +10 vs chain!). Same with the ranged weapons although rate of fire and range differentiates them enough already imo.

    1. I'm glad you saw this! Totally true all around, esp. about the ranged attacks, too. The details of first strike in Chainmail are attractive to me, but I find them hard to work with big D&D groups, so I use just a few echoes of that in my games.

  39. This is a good insight, but is easily adjusted for and still helps to differentiate weapons in interesting ways.

    1) We simplify armor class adjustment factors (ACAs) per weapon by averaging the ACAs of UA across ACs 0-3, 4-7, and 8-10, and redefining these as plated armor, non-plated meatal armor, and non-metal or no armor.

    2) We then correct for the error noted here by adding +2 to the 0-3 range and -1 to the 8-10 range. These mods are debatable but sound.

    The result is that heavy blunt weapons are needed to better hit plated armor, as well as heavy bladed or piercing weapons. The UA Cavalier Weapons of Choice reflect these differences.

    So yes, it is possible to fix the error you found and efficiently implement ACAs, which provide desirable weapon differentiation.

    1. Well, that's a reasonable fix, thanks for that!

  40. Hi Delta, since you asked whether anyone else had noticed this before--the poster Matthew- on Dragonsfoot did back in 2009!


    "The two handed sword modifiers of +2|+2|+2|+2|+3|+3|+2|+1 mirror the pattern 6|6|6|6|5|5|6|7, but their meaning is completely at variance in terms of probability, and the use of variable damage as opposed to a "kill" makes no difference in that respect. Broadly speaking, the pattern should be more like +0|+1|+2|+3|+5|+6|+6|+6..."

    1. Wow, thanks for finding that, impressive work by Matthew! That detail is a little lost in that very long and detailed post. (And I differ a bit a few lines later where he says the "patterns are haphazardly imitated"; my last chart above shows they're identical except in a select few cases.) Excellent find, thank you for that!