This week we’re taking a bit of a detour to critique some video-game armor, in this case the armor of Baldur’s Gate III. I have been meaning to do a general critique of the Dungeons and Dragons 5th edition armor system from a historical perspective for a while, and the massive outsized success of BG3 made this seem like the obvious time to do so. In particular, BG3‘s success, I suspect, will make its artwork the ‘standard’ visual depictions of these armors for many DnD players when they imagine their characters. Moreover, a critique of DnD on this point generally is, I think, useful: DnD remains one of the most common entry-points into pre-modern arms and armor for many people, which has traditionally been a challenge educating in this field, because DnD‘s treatment of historical arms and armor is generally quite bad and its mistakes have a habit of becoming popular ‘knowledge.’
Alas, 5e and Baldur’s Gate III, while they offer some improvements (goodbye, ‘banded mail,’ whatever the heck you were!), still generally display a pretty weak grasp of pre-modern armor, from materials to construction, to fit to function and weight.
But first, as always, if you want to support this project, please share it! And if you want to make sure that I can afford proper armor, you can support this project on Patreon. Finally, if you find this kind of content interesting or valuable, I’d ask that you lobby your representatives wherever you may be to support the humanities in general and history and classics in particular. Shrinking public funding in many countries is causing our disciplines to wither and without the resources provided by robustness in the disciplines of history and classics, this project would not be possible.
If you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings, assuming there is still a Twitter by the time this post goes live. I am also on Mastodon (@bretdevereaux@historians.social) and BlueSky (@bretdevereaux.bsky.social), though I am most active on Twitter still (for now).
In brief, the fifth edition (5e) rules for Dungeons and Dragons divide armor into three categories (light, medium and heavy) and that makes an easy enough division for discussion. In particular I am interested in three questions: do the name and basic rule-book description resemble actual historical armors? Do the statistics and relative weight in turn make sense given the qualities of those actual armors? And finally, do the visual depictions of the armor in Baldur’s Gate III make sense as translations, albeit fantastical ones, of the armor described? I’m particularly interested in BG3‘s visual interpretation because I suspect, given the tremendous popularity of the game, that its interpretation is going to become the standard reference for what armor in DnD looks like for many people.
And since we’re looking at the visuals of the armor in game, that means lots of screenshots, so everyone please welcome our model of the hour, half-elf Oath-of-Vengeance Paladin1 Aimee:
I am mostly going to focus on the base, unenchanted armors; the +1 variants and so on have slightly different designs (indeed, that is a +1 chain mail up there), though in a few cases I am going to comment on the higher variants, particularly when they make the problems of the base variant worse.
Before we dive in, let’s clarify how armor works in 5e. Each kind of armor contributes to a character’s “armor class” (AC), every point of which reduces the chance of being hit by 5% (to a maximum of 95% and a minimum of 5%).2 In 5e, light armors are mechanically distinct in that they allow characters to use their full dexterity bonus3 to armor, whereas medium armor caps this bonus at 2 and heavy armor doesn’t allow it at all. In practice, given the AC values, that means light armor is the optimal choice for a character who cares not at all about stealth at 20 dexterity4 and for a character who can’t afford to be at disadvantage5 on stealth rolls at 18 dexterity.6 While we’re here, plate armor technically beats all medium armors at any dexterity, but is very expensive; at a dexterity of 14 or above, medium armors otherwise match heavy armors.7
In practice the way this works is characters take the heaviest armor their class’ proficiency allows and then, in a point-buy system (and where players are interested in optimal builds), dexterity is tailored to the armor type; light armor classes tend to max it out, medium armor classes cap dexterity at 14 and heavy armor classes will leave it at 10. AC values in 5e (in contrast to earlier systems, like 3.5 which had a lot more ways to build much higher ACs) thus tend to vary within fairly tight ranges.8 Assuming no shields, a light armor rogue might have 15 or 16 AC9 while a medium armor barbarian might have 16 or 17 AC,10 while a heavy armored fighter (wearing splint or plate) might have 17 or 18 AC.11
Light Armor
We start with light armor, which comes in three kinds: Padded, Leather and Studded Leather in order of cost and weight. Studded leather is more expensive, a touch heavier and offers 1 point more AC than the others, while Padded is cheaper but comes with disadvantage on stealth checks in pen & paper. Here is what they look like in Baldur’s Gate III:
And we actually start off pretty strong with padded armor. Padded armor was a real kind of armor and the the rule-book description that this is “quilted layers of cloth and batting” is a decent enough description of how it was structured. Historically, this armor shows up in a lot of cultures and goes by quite a lot of names (gambeson, aketon, padded jack, arming doublet, etc), sometimes worn as a primary armor or worn as foundation layer for other armors.12 At around 3kg or so, armor like this makes a lot of sense in the ‘light’ category too.
As for the visual depiction, I have a few quibbles. Because these are made of fabric, there’s no reason they can’t extend down at least to the knees to offer more protection and for padded armor worn without any other armor layer over them, they generally do. Also – and this is a problem with several of BG3‘s armor designs – there’s no reason this armor needs to split so high up on the body and doing so exposes the lower torso. Instead, an armor like this can extend down to the knees while being a bit loose and closing with an overlap in the center to provide more coverage with no reduction in freedom of movement. Judging from some of the other armor designs, I suspect they split this high up actually to avoid lots of clipping and necessary fabric physics. Finally, the gigantic leather belt the character wears with this armor is unnecessary, though the idea of purely cosmetic, stylistic elements in armor is perfectly reasonable: people decorated armor!
Leather armor creates a lot more problems because the weight-class of this armor doesn’t agree with its rule-book description which doesn’t agree with its appearance. The description says this armor’s chest and shoulder protection is, “made of leather that has been stiffened by being boiled in oil.” That runs into a whole mess of problems; cuir bouilli (‘boiled leather’) is clearly what is intended here, but it is both unclear of cuir bouilli was ever actually boiled and second it seems pretty clear it was not ever boiled in oil. Still, we might assume some kind of hardened leather, but then the problem is that most hardened leather armors are quite a bit too substantial to be in the ‘light armor’ category. Cuir bouilli was a substitute material for iron plates in armor, not for a padded jack; the material is relatively heavy and inflexible.
We’re long overdue to discuss leather in armor here, but briefly the tanned sort of leather in a modern leather jacket isn’t going to accomplish much of anything as armor, while the hardened leathers that will are usually not very light or flexible. However, I can think of one sort of leather armor which would fit quite well into this category: buff leather. Somewhat flexible (though less than your leather jacket), coats of buff leather (“buff coats”) were used as protection underneath armor and on their own in early modern Europe and became something of a signifier of aristocratic status in artwork.
The visual depiction splits all of these differences in unsatisfactory ways. It has a light shine to it that makes me think they were thinking of cuir bouilli, but the light-tan color is not generally a color one sees cuir bouilli in (though it isn’t an impossible color by any means), but the thickness an structure isn’t consistent with cuir bouilli, which had to be arranged in much the same way as metal plates: either as smaller scales or lamellae or as fitted plates like in a plate harness. Meanwhile, they don’t have the very distinctive yellow-orange color of buff-leather, or its normal shaping. Personally, I’d love to see the ‘light’ leather armor of DnD traded out for buff leather, with heavier cuir bouilli armor showing up in higher categories.
All of which now brings us to studded leather, a standby of medieval fantasy, but unfortunately not a thing to actually existed. The added protective value of a few rivets run through leather is basically nil. Instead, ‘studded leather’ seems to be based on a misinterpretation of medieval artwork, what I am going to refer to broadly as the ‘English Effigy Problem.’ What I mean by this is that modern folks (often Victorian moderns) looking at statues (famously, English coffin effigies, but also manuscript illustrations) without any experience of the objects being depicted tend to badly misunderstand what it is they are looking at.
In this case, there was a kind of armor in the Middle Ages which, from the outside, resembled a leather or textile coat with rivets through it: the brigandine (and also some coats of plates). But those rivets weren’t the defensive element, they were securing many small overlaping metal plates and those were the defensive element. A brigandine was not by any means light; it was a viable substitute (and with the coat of plates, a chronological forerunner) for a full breastplate. Looking at the BG3 armor, even this many years removed from the original mistake of having ‘studded leather’ in DnD, it is still immediately clear to me that this is a misinterpretation of a brigandine. Studded armor wasn’t a thing.
Medium Armor
On to medium armor.
The first armor in the medium category is hide armor and in the rules it is explicitly an armor of “folk who lack access to the tools and materials needed to create better armor” consisting of “thick furs and pelts.” And while there are certainly armors made of rawhide (a kind of hide product), ‘hide armor’ of this sort was not a historical armor. This sort of thing seems to me to derive from Hollywood depictions of the armor of Mongols and other Steppe peoples, but they frankly had access to much better armor options. The visual depiction in BG3 further seems to confirm to me that this is intended to be armor just made of some thick pelts and furs thrown together, rather than something intentional.
That said, hide occupies a position where a hardened leather armor, made of cuir bouilli or rawhide would make a lot of sense! These hardened kinds of leather, thick, stiff and relatively heavy, were used as armor materials. This sort of armor tends not to survive, but we know, for instance, that lamellar armor – a form of armor made of small overlaping metal or hardened leather plates – was often made with hardened leather when metal could not be afforded. An armor of hardened leather scales or lamellar would thus make a lot of sense as the cheapest form of medium armor.
Next up is the ‘chain shirt.’ This is a real historical armor, though I would call it a ‘mail byrnie’ rather than a ‘chain shirt.’ ‘Chainmail’ is a modern term that historians generally won’t use; mail isn’t made of chains in any case but of linked armor rings. Where I think this goes a bit wrong is in how DnD and BG3 understand a mail shirt to be worn. In particular, the rulebook description says, “a chain shirt is worn between layers of clothing or leather,” which muffle its noise. Now mail was always worn over some kind of foundation garment (like an arming jack), but it was very often worn without any kind of cover over top and certainly not something heavy enough to muffle the sound of the links as one moved.
That problem is intensified with the BG3 artwork. The problem here, I assume, is the desire for a level of ‘visual interest’ in the armor, but mail generally looks pretty simple.13 So instead the mail is actually set into a leather backing, which ties up in the center with a wide leather fringe. Now there’s no reason you couldn’t sew mail into a backing like this, but apart from the shoulder-guards of very early mail, it seems to have almost never been done. The mail shirt was a separate garment, put on after the arming jack. Mail ‘voiders’ (bits of mail to over gaps in another kind of armor) might attack to an arming doublet, but the chain shirt isn’t a set of voiders; you’re not wearing anything over it. So while this is a real armor, it doesn’t really look the way it is shown in game, though the depiction is not entirely impossible, so long as we assume that the mail continues beneath that textile (otherwise there are huge gaps in the armor over vital areas).
Next up is scale mail. So I should note that the addition of ‘mail’ to the name of armors to make ‘plate mail’ or ‘scale mail’ is wrong; there is scale armor, plate armor and mail armor (armor of interlocking rings). But the basic concept in the rulebook is a real kind of armor, consisting of a backing – the rules naturally specify leather but textile could be used – with metal or hardened leather scales attached. Scale armor is distinct from brigandine or lamellar in that the scales are attached to the backing at the top, rather than to each other at multiple points; that makes it easier to get a weapon in underneath the scales of scale armor, making it a simpler but more vulnerable form of armor compared to brigandine or lamellar. And this makes a lot of sense as a ‘medium’ armor, though I should note that a coat of scales could weight every bit as much as a breastplate or a brigandine; there’s no less metal there, it’s just not as solidly attached.
But the visual depiction here is an absolute mess; I’m not even fully sure I know what I am looking at. It seems like there is a base layer of scales, but then leather straps that go over those scales and then also leather (scales?) running down the center (which is exactly where you would want your more durable metal scales!). This doesn’t look like any scale armor I know. It’s odd: game designers tend to overuse scale armor because it is more visually interesting than mail, but apparently here even that wasn’t enough and so we end up with a wildly overworked scale armor design. Scale armor is conceptually simple: you have a backing material and then you layer scales over it. And then you are done! Adding more layers is going to prevent the scales from moving, which will surrender the one major advantage of scale armor which is that it is flexible!
Next up is the ‘breastplate,’ and once again while this is a real armor, the artists have somehow managed to get it quite wrong. The rulebook description also, frankly, bothers me, describing this as, “a fitted metal chest piece worn with supple leather.” At this point, our leather is getting properly out of hand. You could wear a breastplate over something like a buff coat, but it would have been far more common in the sort of societies we are talking about to wear it over a padded textile arming doublet. There does seem to be a trend here where the DnD rules feel like every bit of flexible material has to be leather because leather is strong and badass. But there’s a reason no one wears a leather T-shirt: it’s uncomfortable. You don’t usually want a lot of leather, no matter how supple, right up against your skin. You want cloth. So base layers tend to be textiles.
And then the artwork. The BG3 artwork for the breastplate is a mess. The breastplate is a tiny plate, connected to a tinier backplate by straps, with a whole second ‘belly guard’ below it. This is not what a breastplate looks like. This is what it looks like:
And that basic form is not new, here is a breastplate from 600 BC:
That’s how a breastplate is structured: it extends from the shoulders to the natural waist (not the beltline but also not halfway up the ribcage!) with a front- and back-plate that connect directly together, providing good protection for the sides of the body and usually featuring a downward projection at the base which flares out to allow movement while protecting, to at least some degree, the pelvis (if this part is very substantial it is called a ‘fauld’). There’s no reason to have a second belly-guard when that can just be an incorporated part of the main breastplate. Likewise, there’s no reason not to have the cuirass wrap around and join at the sides, thus removing the vulnerability there.
Finally, the garments underneath the armor are a mess. The character wears a quilted green garment that I assume is meant to be a padded jack (good) but over a second quilted ‘leather-brown’ garment which is…I don’t know? Is it another padded garment, but faced in leather? If so, that wasn’t the normal way to make them. Also it covers the legs; these sorts of things generally did not extend over the legs. The whole thing gives the real impression that the artists understand that medieval armor sometimes involved quilted garments, but that they haven’t yet sussed out when or why, so they’ve just put quilting patterns on everything. Meanwhile the outfit has knee-guards (poleyns) but no greaves or any other metal protection for the legs or arms, which isn’t an impossible configuration, but it is an odd one.
Breastplates should not be this hard; the Pillars of Eternity games absolutely nailed this design already.
On to the half-plate. I have actually found folks surprised when I tell them that of all of the armors in the DnD system, half-plate is not one of the ones simply invented out of whole cloth. Half-plate was a real thing, although we probably ought to say ‘half-armor’! This was a term used to describe late medieval ‘munition’ armor, also called ‘Almain rivets.’ The rule-book description isn’t way out in left field, but also isn’t a particularly good fit: “Half plate consists of shaped metal plates that cover most of the wearer’s body. It does not include leg protection beyond simple greaves that are attached with leather straps.”
In practice, half-armor consisted generally of a breastplate with a backplate, generally with long tassets (segmented, articulated metal plates covering the upper thigh), a metal collar and spaulders covering the shoulders and upper arms, providing very strong protection for only the most important parts of the body and avoiding all of the really hard to armor areas. That gave this armor the advantage of being cheap, modular and one-size-fits-most, and it tended to be an armor stockpiled for issue to regular soldiers in the very late medieval and early modern period.
But the BG3 artwork treatment here is baffling. The historical original is actually a visually interesting armor, especially once you put it over an arming doublet and add some details! But the BG3 artwork has instead gone with an armor, again, where I have a hard time figuring out exactly what is going on. Once again, the breastplate is inexplicably literal in covering only the upper third or so of the torso instead of extending down to the natural waist and below it we have what looks to be three layers of alternating leather-textile-and-leather? And then we have what looks to be at least four belts.
All in all then, medium armor from a rulebook perspective is alright, but the visual choices for BG3 are almost all quite a bit off; given the tremendous success of BG3, I suspect this will exert a significant ‘pull’ in how players imagine and understand their medium armor, which may counteract the somewhat better rulebook descriptions. On to heavy armor.
Heavy Armor
The first of these is ring mail and this is an easy one: ring mail is not a thing. The rules say, “This armor is leather armor with heavy rings sewn into it” and that’s just not an armor anyone ever seems to have used at any point, historically. ‘Ring armor‘ like this was conjectured armor during the Victorian period, composed mostly of misunderstandings of artwork. Visually, the structure of the armor in BG3 reminds me quite a lot of a brigandine, in terms of the length of the coat and its front-joining, leading one to wonder if this is another example of the ‘English Effigy Problem.’
The design for ‘chain mail’ is much better. The rule-book description even specifies that mail gets, “layer of quilted fabric worn underneath the mail” rather than leather! And the core of what we see in the game at least has a decent resemblance to a mail hauberk. And, for visual interest, they’ve added a surcoat, which is good, mail was often worn with surcoats as decoration and for temperature control, so a decorated surcoat is a great addition. There are a few problems here, though. The first is in the ‘cut’ of the mail, which includes split sides and a large front opening; for mail this long, some split is necessary, but it was usually just a front-and-back split at the base of the mail and it didn’t extend nearly this high. More frustrating are those hardened leather pauldrons and collar; that’s not an impossible configuration, but it wasn’t a particularly common one either and surely if one was going to add reinforcement, they’d reinforce the chest first?
Looking at the +1 version clarifies this problem as the quintessential ‘English Effigy Problem.’ We’ve now added metal pauldrons and a high collar that is almost a bevor protecting the neck, while keeping the rest of the armor simply in mail with a surcoat. I’ve often supposed that the problem here is a misinterpretation of some medieval sculpture (again, particularly certain English funeral effigies depicting figures in armor) where the figure appears to be wearing a lot of leg, arm and neck protection, but just what might be interpreted as a surcoat over mail over the body. But what that figure is actually wearing is either a coat of plates or a plate cuirass covered with textile (such as a jupon).
I also find I have an issue with placing mail so far up the weight scale as to put it in the ‘heavy’ class. Certainly, a full maul harness (which might have included with the hauberk full sleeves, hand protection, and leg protection (chausses)) could be fairly heavy and was in its day the heaviest armor available in Europe. But in a setting where full plate protection is an option, a simple mail coat like this was no longer ‘heavy’ protection. And the rule-book weight of 55lbs is wildly off; a heavy mail coat of this sort might weigh 30lbs, tops. Most are closer to 20-25lbs. The rulebook cost is also comically wrong, with the ‘chain mail’ costing 75 ‘gold pieces’ to half-plate’s 750 (remember, half-plate was a cheap armor for cheap soldiers) and the breastplate’s 400. In fact, a mail hauberk would have been much more expensive than either. In particular it’s strange to put the mail hauberk in the same weight-class as plate armor, which historically often incorporated a complete mail hauberk as part of its foundation.
After this comes splint armor and the historical interpretation problem is clear. Splint armor, which is, as the rule-book says, “narrow vertical strips of metal riveted to a backing” of some flexible material, was never a complete armor, but a form of protection for the arms and the legs, with the body being protected generally in some other way (like mail, or a coat of plates). But the phrase ‘splinted mail’ is coined by Victorians looking at armors which incorporate splint elements into the arm and leg defense, leading to the assumption that this was a complete and distinct kind of armor, rather than just a specific way to make rerebraces (upper-arm guards), vambraces (lower arm guards) and greaves or cuisses (leg guards).
The in-game visual depiction is also a mess, with a tanned-leather base with a mix of lamellar that doesn’t overlap vertically (which defeats much of the purpose of lamellar) and segmented plates which also don’t appear to articulate; the whole thing ends up looking like someone threw three or four different armors in a blender and took what came out, but I can’t fault the artists too much because they’re being asked to imagine how a design principle (‘splint’) could be applied to parts of the body it was never applied to. Nevertheless, I can’t help but notice that the BG3 ‘splint armor’ includes no splints on the arms or the legs (or any other part of the body).
In practice, I think this would have been an excellent spot to simply have a lamellar armor with metal lamellae or alternately a Roman-style segmented plate armor. Another option here could have been some kind of reinforced mail armor, as we do see societies where instead of wearing plates over mail, metal plates are integrated into mail.
Which at last brings us to the end with plate armor. The basic rule-book description is fine, “shaped, interlocking metal plates” covering the whole body with “thick layers of padding underneath.” One crucial thing missing here however is mail; plate armor was generally worth with either a mail shirt between the padding and the plate or with mail ‘voiders’ – smaller pieces of mail made specifically to cover the gaps in the plate protection around joints.
Still, this is a real kind of armor and was the heaviest and most complete sort of protection we might expect in a late medieval society. The BG3 artwork here is colorful and detailed and works hard to sell the, “This is really expensive!” part of concept, but leaves some things to be desired, at least with the base version. The aversion of the artists to fauld-and-tassets means that there are gaps between the leg harness and the breastplate (half-obscured by an oversized cloth skirt, another feature – belted cloth skirts – the designers love which were not, in fact, common in these sorts of armor) and the leg harness as a result runs very high; it looks like bringing the leg fully up might be pretty uncomfortable. Meanwhile, there are visible gaps between the spaulder/rerebrace (upper armor protection), which is normally where a pauldron would go (or mail voiders!). The breastplate also has a multi-piece riveted design which we might say is for artistic flourish, but actual breastplates often avoid that sort of thing, preferring instead a globular shape to better let blows glance off. Finally, the cuirass has the same problem as the others, where the breast and back-plates don’t fit together well, leaving a visible gap where none is needed in a properly made armor.
Closing the Gap
So how might I ‘fix’ this system? I think its worth thinking in two ways: what might we do while keeping the basic mechanical structure (light/medium/heavy and the AC system) in place, and what might we do if we weren’t confined to that system.
For the former, first I think we can safely remove the armors that are made up: studded leather, hide, ring mail and splint can all go.14 We replace ‘leather armor’ with ‘buff coat,’ at roughly the same position in the set. We can then add in some of the transitional armors oddly left out: brigandine and lamellar, to fill in the gaps created by getting rid of the made-up armors. Next, some re-ordering, with our conceptual division being less about pure weight and more a function of restriction of movement and fatigue. Our basic mail byrnie (replacing the ‘chain shirt’) moves to light armor (back where it was in 3.5, I might note), while a mail hauberk becomes the standard medium armor (again, back to 3.5). Meanwhile our heavy armors are reserved for fully rigid armors covering most of the body. That gives us, in roughly ascending order of AC:
- Light Armors
- Padded
- Buff
- Mail Byrnie (‘Chain shirt’)
- Medium Armors
- Mail Hauberk (‘Chain mail’)
- Scale Armor
- Hardened Leather Lamellar
- Brigandine/Metal Lamellar
- Heavy Armor
- Breastplate (over mail)
- Half Plate
- Full Plate
Now what if we could completely change the mechanics?
The biggest mechanical change I would make is this: armor in a real fight doesn’t generally negate damage the way armor class functions in DnD. A strike which hits armor rings or a breastplate isn’t cancelled out, but merely converted into blunt-force trauma. Consequently, I’d want different kinds of armor to bring damage reduction rather than a set chance to negate incoming attacks. Armor class is a huge part of DnD’s balance, so it would be hard to remove without overhauling everything, but I might also give escalating damage reduction (something like DR/magic in ye olde 3.5 rules). To go even further I might add a rule that all hits which beat the AC rating always do at least one damage.
What would I do if I could rip out the whole system and start over? Well, I run my own games with a heavily homebrewed version of the d10 (‘Storyteller‘) system, in which armor purely serves as damage reduction. Each character can have one type of armor in each of three layers, which approximates the layering of medieval armors in Europe and the Middle East: a padded lower layer, an intermediate mail layer and then a rigid armor surface layer. The ‘soak’ (=damage reduction) values of each layer are added together to give the character a single, fairly simple to manage armor value. Since the Storyteller system splits damage types into bashing, lethal and aggravated, that gets expressed as a soak value against each damage type, with armor downgrading each level of damage into the next lowest (aggravated becomes lethal, lethal becomes bashing, bashing becomes no damage at all). So a fully plate-armored knight might have a 1B/4L rating, meaning the first four points of lethal damage in an attack are instead treated as bashing and the first bashing point is negated.15
The mechanical result is that players facing an armored opponent either need to score a big enough hit to overwhelm the armor or find a way to attack that avoids it.16 That in turn creates a balance space for weapons designed to specialize against armor, something that DnD largely lacks, but which was a significant concern historically.
- For some reason, when I planned this, a part of my brain thought, “oh, I’ll need to pick a class with all three armor proficiencies. In retrospect I did not need to do this just to take screenshots, but, oh well, paladins are neat anyway. Also the added base moment speed half-wood-elves get is really handy with a melee-focused class like paladin.
- For those familiar with the game this may seem an odd way to sum this up, but this is the statistical impact of the dice system.
- (Dex-10)/2, rounded down
- Where studded leather will result in 17 AC, a statistical tie with the heavier, more expensive half-plate.
- Forced to roll two dice and take the lower result
- Where studded leather statistically ties a breastplate at 16 AC, while being lighter and cheaper.
- Half-plate at 14 dex or higher gives 17 AC, equal to splint and greater than chain mail or ring mail. All heavy armors give disadvantage to stealth, so that’s not a factor.
- Most non-armor classes have access to something like Mage Armor, which gives them a base AC very similar to armor-wearing classes. Mage Armor sets a caster’s armor at 13+Dex Bonus, making it equivalent to a +1 Studded Armor, so long as the caster doesn’t wear any armor.
- Studded leather with 16 or 18 dexterity
- 14 or 12 dexterity with either a breastplate or half plate producing a range of 15 to 17, depending on build
- Though at low levels characters might not be able to afford optimal armors. Assuming cost-constrained characters, we might expect light armor at 16 AC, medium armor at 16 AC and heavy armor at…16 AC (Studded, Scale and Chain mail respectively). And the wizard with Mage Armor might have around 15AC (assuming 14 dexterity). It really does vary within a relatively narrow range, especially when you remember that 1AC is just a 5% chance to hit; significant but not enormous.
- Structurally, I should note that arming doublets worn under armor might be differently constructed (in terms of thickness, facing materials, etc.) than padded jacks intended as a primary armor.
- Though decorations to mail armor, like incorporating bronze or gilded rings to add color in patterns or as fringes, were common enough.
- Splinted arm and leg guards could remain as components of other primary armors.
- The base storyteller system allows character to further negate levels of bashing damage with a stamina roll, so a knight with a lot of stamina has a good chance to negating the rest of those bashing damage levels.
- The base storyteller system includes rules for taking a higher difficulty on the attack role to target specific body parts. I ended up codifying these by giving certain weapons higher difficult ‘precision’ attacks which lessen or ignore armor, making weapons historically designed to defeat armored opponents dedicated ‘can openers.’ Alternately, I set the base damage on some blunt trauma weapons high, making them good choices since they bypass the lethal damage protection.
A shame, nothing about shields? Those were quite an important piece of the defensive toolkit, maybe even the single most important piece for a lot of fighters over a lot of history, and I’m a little surprised not to see anything in them here. I always had a vague notion that (at least as of 3.5, which was my last foray into dungeons and dragons before moving on to other systems) the Tower Shield + 1 handed weapon + full plate that many low dex fighters and paladins went for was simply never a thing historically, and that the extra protection of the shield was most necessary for people wearing lighter armor.
Also, because I’m feeling nostalgic and it’s a shameless plug, there’s this very old system Dragonquest. It worked off of a damage reduction model, with different armors having a different amount of DR but the extra weight actually lowering your agility and making you a bit more likely to take hits. But there was also a mechanic that you could get a ‘penetrating’ hit if you rolled very well. (exact parameters got a bit complicated but I could go through the math if anyone actually cares). Penetrating hits ignored armor, and something like full plate had a hefty 8 DR, which could and often did completely block a weapon, meaning heavily armored fights could be a bunch of maneuvering around trying to set up your attack bonuses that you had a decent shot of getting a hit going through the armor, which was somewhat necessary to compensate for not having a dedicated ‘precise attack’ sort of option.
This post is 6,200+ words long and you are surprised not to see it tackle an extra subject? Have mercy on what little remains of my free time!
The post is over 6,200 words? It doesn’t feel it. I could have sworn I read it in just a few minutes.
My apologies for the thoughtless remark.
Same here: this is a remarkably easy read. Partly, the professor writes well; partly, by the nature of the topic the bulk of this post is structurally similar to a listicle.
On the topic of plate-armor-and-shield I recommend the Lannister and Gondor infantry kit review posts.
So one issue that makes things confusing is that “damage” and “hp” are very esoteric concepts in DND. Here is an excerpt from Gygax’s definitive statement Much About Melee:
“Hit points are a combination of actual physical constitution, skill at the avoidance of taking real physical damage, luck and/or magical or divine factors. Ten points of damage dealt to a rhino indicates a considerable wound, while the same damage sustained by the 8th-level fighter indicates a near-miss, a slight wound, and a bit of luck used up, a bit of fatigue piling up against his or her skill at avoiding the fatal cut or thrust. So even when a hit is scored in melee combat, it is more often than not a grazing blow, a mere light wound which would have been fatal (or nearly so) to a lesser mortal. If sufficient numbers of such wounds accrue to the character, however, stamina, skill, and luck will eventually run out, and an attack will strike home…”
From Gary Gyax’s article Much about melee published in Dragon #24 (April, 1979).
Of course this still isn’t perfect but it does explain a lot about some weird aspects of DnD. We can also probably conclude that many later DnD designers did not understand this. Certainly most players, even avid DnD junkies, are often unaware of this context. While the visual/flavor types of armor are not much impacted by this it does help explain the mechanics of the system to some degree.
I don’t know that later D&D designers have necessarily forgotten HP is supposed to be abstract. I know the 5e books are very explicit about it. The problem is that “what do HP represent?” has always been a tension in D&D, going back to Gygax’s version of the rules, because on Gygax’s account it makes no sense for a healing spell to heal 1d8 HP regardless of whether it’s being used on a rhino or an 8th level fighter, nor does it make sense for Constitution to benefit HP while Dexterity benefits AC.
Under the argument advanced in the quote above Armor Class *is* damage reduction as Bret explains it. That’s my point.
It also does make sense for a healing spell to be flat health and not a percentage. The power of the spell applies to the abstraction. In some sense you are counting stuff like stamina/vigor as part of health. Many spells like “Vigor” provide minor healing, and Vigor is really more like anti-exhaustion.
Also in OD&D Dex originally only impacted to-hit with missiles and later it impacted melee but only for fighters.and the first 3 scores, Str/Int/Wis only impacted exp bonues/penalty for fighting men, magic users, and clerics.
It is notable that DnD wasn’t a college dissertation or something. It won’t stand up to rules laywering or even “concept lawyering”. It was a whimsical alternative to Chainmail that initially didn’t have a lot of thought put into it.
The thing to remember about about 5E books “being explicit about HP as an abstraction” is that it doesn’t mean that the people designing new versions of the game shift the mechanics significantly based on the flavor they are using.
It’s not surprising that an anti-exhaustion spell would restore HP. It’s surprising that it restores a flat value, if the flat value can represent such different scales of wounding. By the Gygax quote earlier, a vigor spell can energize an experienced warrior, and no more – or close a mortal wound on a rhino, or a civilian.
That’s only a problem if you treat the magic as strictly consistent and rather unmagical. There’s nothing in the rulebook that says that casting cure light wounds and on a level 1 character and a level 20 character are the same thing. Whereas a 5 hp wound is a serious injury for a level 1 character, for a level 20 character you are talking about a demigod being marginally drained. Is it surprising that adding a bit of healing energy to a demigod wont be as noticeable as a random peasant?
Unmagical?
I mean, I find the notion of “that is not what magic is like” rather odd given that magic neither exists nor is its possibility extrapolated from things that do exist. Could you explain, please?
(My current frame of view is that is magic existed, then — tautologically — it would be just another part of the natural world, and people would relate to it as they do to the rest of the world. Which is to say, just as ancient societies made practical use of various phenomena, based on accumulation of trial and error, and surrounded that use with “religious” rituals that had no pragmatic effect on the phenomena, they would likewise make practical use of magical effects and surround such use with “religious” rituals of no influence on magic. By the same principle, in our society magic would be studied, and after a few decades its basics would be taught in school. The same teenagers who are bored by having to learn e.g. chemistry would also be bored by having to learn magic. And just as every so often, people accidentally make chlorine gas by mixing hydrochloric-acid-based and sodium-hypochlorite-based cleaning agents, every so often people would create small magical accidents by mixing cleaning agents with “incompatible” magical properties. Is this frame incorrect?)
Considering that chemistry encompasses a large portion that was seen as magic by the ancients, you might make a solid argument that our current science is the systemization of magic you describe.
Exactly! Magic is unexplained causality. Stick it through the wringer of the scientific method, and it comes out either Science or accident.
Of course, the usual fantasy approach is it came through the wringer differently.
You might want to try Ars Magica.
If you haven’t read it, you’re gonna love it!
Usually “hp is an abstraction for stress + luck + exhaustion + physical wounds” is brought up in order to argue that high level characters are not superhuman. The HP of a high level character is high not because his skin has become unbreakable and his gut can absorb bullets. No, he’s as mortal as any random schmuck, just far more skillful and always manages to place something between his neck and oncoming blades.
When you say that a level 20 character is a demigod whose vitality is just too great for a “cute wounds” spell to do much to, that’s a possible interpretation. But it’s an interpretation that concedes what the abstraction argument is trying to push back on – you are agreeing that yeah, a level 20 warrior is not really human anymore. The 300hp barbarian is really a minor tarrasque in a much denser body. He fights by being stabbed in the eye and having the rapier break before his inviolable eyeball.
You’re conceding it the second you decide ‘cure light wounds’ heals 1d8 damage, because that implies 1d8 damage is a light wound.
This is what you get when a guy cribs a bunch of mechanics from naval wargames and tries to use them for dungeon crawling without bothering to see if they make sense for people who are not 200m long and made of steel.
So what you’re saying is that the rules make more sense if you imagine the D&D universes are populated entirely by shipgirls.
And all the others beside, as previously alluded to on this blog. That damage mechanism (including “the cringe of systems where the damage stat belongs to the arrow instead of the bow“) makes perfect sense for projectiles carrying an explosive filling.
Or from WW2 squad tactics. You can figure out a military system with D&D physics, in which magic means that line infantry tactics do not work (a pike square just screams “fireball me”) and thus war has to be fought with the modern system. But that requires being a lot more careful about social, military, and social-military history than D&D has been so far.
Half the fun of reading kingdom building litrpgs is laughing at scenes where people with ridiculous super powers are trained by the modern world MC in advanced post classical military tactics that don’t make sense when people have super powers.
I mean if you want that you can look at errebon. I don’t think the setting focuses on it since it’s in the backstory but that was a setting where large nations going into a giant war plus those nations having magic that put them around the industrial revolution would probably lead them to behaving like armies during ww1. They even have magical equivalents to things like artillery or mines.
The other problem is that realistic pre-modern small-group combat isn’t fun. Nobody wants their lovingly crafted PC to get randomly speared in the face and die at the begining of a fight.
Depleting some scalar quantity (which does make much more sense if it represents the number of soldiers in a formation left alive or even the general integrity of a warship) is a blunt instrument but a sensible one if you want to give combat non-lethal consequences.
Though if that’s what you wanted, Runequest and GURPS would accommodate you. But of course those systems never achieved a fraction of D&D’s popularity.
Yeah, so D&D physics has an answer: easy healing. The upshot is that the winning side in combat will not take any deaths, because the 0 hp characters can be healed. This means that winning the battle takes priority over casualty avoidance, which leads to very different tactics, as well as a very different military culture, more first system than second system.
It sort of makes D&D very good for getting you a proper barbaric epic feeling: the combat is about winning, and your characters get to do great deeds, just like Beowulf or Hector. Essentially, the mechanics encourage you to fight like a character from Icelandic sagas, without too much care for personal safety. Yep. The first system it is.
I played a home-brew for years where HP was low, and any roll that hit was followed by a roll ‘To Kill’ (effectively out of combat). It kept fights short (and put a premium on deciding when to run away). On average PCs died about 5 times before reaching medium to high level. No-one complained, although a few tears when loved characters died.
Reminds me of, I think, Alan Zimm talking about wargaming naval combat in the context of Pearl Harbor and making the point that it’s very important whether you decide to model damage as “this battleship has 240 Hit Points, each round on target reduces this by a certain amount depending on the ordnance used, and when those are expended it sinks” vs “each round on target has a certain chance of inflicting a critical hit, and that chance is higher for larger ordnance”. If you’re using the wargames to train your officers and plan your attacks, as the Japanese did, this will make an absolutely crucial difference.
Let’s say combat experience has told you that eight hits with a 6″ shell will, on average, sink a cruiser. If you are fighting four cruisers and you know you’re going to be able to land 16 hits on target during the engagement, then under system 1 it makes no sense to do anything other than focus on two of the cruisers and ignore the rest. It takes eight hits to sink a cruiser, it will always take eight hits to sink a cruiser.
But if each hit instead has a 1/8 chance of sinking, well, then, it’s worth rolling the dice a bit, because you might get lucky and sink all four…
Yeah this “HP as a potpourri of everything keeping you on your feet from your health to your luck and everything in between” is fine in theory but once you throw healing in it all stops making a lot of sense. Much like alignment is kinda thrown out of wack once you have it be an actually measurable thing with mechanical effects
I question how well it works even before healing is factored in. Look at what influences your HP; your Constitution score (physical hardiness), level (training/experience), and class (whether you’re a martial jock or a spellcasting nerd). Evasion, luck, magical/divine factors, and so on don’t factor into it, except arguably in character level.
All of that feels like post hoc justification for mechanics designed with only gameplay in mind. I prefer when they just go full anime and say “Yes, your mid-level gnomish paladin can survive attacks that would fell an elephant, is that a problem?”
Yes they do factor in. It literally says that in the quote from Gygax. Also some of that stuff is part of armor/hit class.
“once you throw healing in it all stops making a lot of sense”.
Not if you look at healing as a retcon that fully defines the nature of the damage. The character got hit for 10 points of damage, and is now down to 20 HP. The number is an abstraction, but the form the damage takes depends upon the method used to heal it.
If you use Second Wind, I guess you were only bruised and worn down from that sword blade you barely parried.
Get an actual healing spell cast? The blade lacerated your arm and the spell healed the wound back up.
Waited until a Long Rest to restore all of your HP? You were only weary from a day of battle.
Alignment is taking every moral issue that the best and wisest have broken their hearts over for millennia, misunderstanding half of them, boiling them down to a rule set, and handing them over to gamers many of whom are sophomoric (and only some with the excuse of being sophomores.)
That was terrific!! My browser won’t let me Like this, but I tell you directly I Like this and am copying and saving it to incorporate into my D&D books. Thanks!
Glad you like it.
When I have a bit of time this afternoon I will see if I can align different schools of jurisprudence with the 9-box grid of DnD alignment.
Yes, you can. The question is whether it can be done well. You see all sorts of memes with rather — dubious — choices because someone was determined to fill it up using a theme.
I maintain that the best explanation of hit points is that they’re a quantification of how anti-climactic it would be if this character died right now.
Pretty much, yeah.
I’ve seen and heard plenty of arguments about how hit points are dumb and don’t make sense, but when someone actually writes a replacement it always comes out pretty complicated. And if there’s one thing D&D doesn’t need it’s to become more complicated.
I do treat them as “plot points”
Why did Minions in 4th Edition only have ONE HP, regardless of how beefy they were (you could theoretically have Minion Dragons)? It’s because they’re not important to the plot, so when they get hit, they go down.
It’s interesting to compare this to the Blades in the Dark system which inverts the what is represented. In DnD the injuries are abstracted into the GMs narrative but the resilliance is explicit in the form of HP. In BitD there are no hit points but the injuries are explicit, if you’ve got a broken arm it doesn’t matter how experienced you are, your arm is just as broken and the same mechanical disadvantages will apply. In theory these systems seem to diverge but in practice the results are convergent. Sure there aren’t any hitpoints to say that the same attack will do different damage to the greenhorn and the veteran of 20 adventures. But the BitD tells GMs to make injuries what is narrative appropriate and over the course of a campaign, that’s naturally going to evolve even without any explicit mechanics. So as long as you keep your eyes on the narrative, it does the same thing.
I think DnD gets a lot of critique from people who forget that the rules are supposed to serve the narrative not the other way around.
D&D grew out of the appendix in Chainmail, a miniatures rule set intended to simulate battles in a fantasy milieu where the armor and weapons were more or less medieval, and the magic was Vancian (i.e. it was a kind of mathematics instead of appealing to spirits to curse your enemies).
The ruleset included commanders and heroes, individual figures that had effects on battles, both attached to a unit and even solo.
The appendix had rules for playing with solely the commander/hero figures, and what they would do. Importantly it included rules for growing a figure from a tyro that would be in a pike line to one capable of dominating the field like in the main set.
The rules were intended as *simulation* – of course of a fantasy milieu rather than of any real world,
There’s been a lot of things grafted on to the rules in the service of narrative, e.g. lots more hit points based on constitution score, negative hit points, death saves, every character having healing on rest, etc. All to let characters’ stories continue (and to avoid having to make up an 8th level character from scratch in the pizza break because the dice had it in for you that day).
Other, pure narrative-based rulesets came later – I suspect from people who decided that a fresh start was a better approach.
The problem is that it’s impossible to arrive at a really self-consistent understanding of hit points in original D&D or its descendants, whether you think HP is 1% meat and 99% grace, 99% meat and 1% grace, or anything in between.
1 damage to a 100-HP creature is, necessarily, actual physical harm because it can have ‘riders’ like injecting poison from a poison-tipped dart. If the dart didn’t pierce the flesh ‘on hit’, it wouldn’t have been able to convey the poison. Thus HP is almost entirely ‘meat points’.
And yet, even 99 damage to a 100-HP creature doesn’t seem to involve any actual physical harm because it has no immediate in-game consequence on the creature (except in 4th edition, and even then the effect isn’t “a massive reduction in the wounded creature’s physical capabilities” like you would expect). Per rules, the creature doesn’t fall over or involuntarily scream or pass out from pain or anything. So HP is almost entirely ‘grace points’.
So, contradiction.
Finally, suppose 100 damage is exactly enough to kill some creature. (i.e., 100 HP in OD&D, 100 minus X HP in latter editions). For that creature, 99 damage is as close to a mortal wound as you can get. In modern editions, this damage goes away overnight! In contrast, I’ve had papercuts that took a week to heal. And even in OD&D, healing cannot take more than a month. That would be optimistic for being discharged in a modern hospital after “a wound that very nearly killed you”. It would be the beginning of years of physiotherapy.
There are two ways to deal with the armor issue. First, as you note, is just to relabel the armors to better correspond to historical categories. This is akin to the DM technique of ‘reskinning’ a monster to make it more a little more interesting or to disguise its capabilities from seasoned players. (For example, turning all of the kobolds in your adventure into ‘batoi’, little bat-people who scurry along with vestigial wings.)
When we get into your proposals for mechanical changes, however, we have to ask, “What are hit points?” And this is a space where D&D has been confused from the very beginning. Are hit points damage? Are they fatigue? Luck? Some combination of the above? You can go back to the very earliest of Gygax’s writing and see all of these explanations and then some laid out as justification for what hit points represent.
In my games I prefer to see hit points as some combination of fatigue and minor wounds/bruising. (Which, of course, contradicts the ‘cure wounds’ spells to a certain extent.) From a mechanical perspective, then, AC negating hits entirely makes a great deal of sense. Given it’s dual role as avoiding hits (DEX bonus!) and attenuating impact (heavy armor!) then it’s a fine mechanical tradeoff.
Ultimately, hit points in D&D are ‘good enough’, especially when we’re talking about action movie representation. Remember that most RPG players aren’t looking for ‘realism’ (though they may say so.) What they really want is verisimilitude; they want it to feel like the right kind of movie or novel. Right now, D&D 5 is simulating heroic medieval fantasy action movies, and hp-plus-AC simulates that quite well.
I think hit points (in general rather than D&D specifically, and in systems aiming to have an actual degree of simulation – level 1 World of Warcraft character having 100hp and level 80 one having 40k doesn’t have any mapping to reality of the setting) are best thought of as “the maximum amount of punishment a body can take”. Some wounds are immediately incapacitating and lethal, some cripple certain functions like the use of the wounded arm, but a lot of wounds are neither immediately lethal or incapacitating: I don’t know to what degree this is a literary trope and to what degree it’s just how historical combat used to function, but historical sources often talk about combatants fighting on despite their numerous wounds, or felling after receiving many. That is to say, having your head cut off kills you outright, but a combatant finally succumbing after 10 strikes that successfully penetrated armor but didn’t strike the vitals can very well be abstracted by each strike dealing one point of damage, and the combatant having 10HP.
I have recently been wondering about this “dying after receiving several non-lethal wounds” (for a game idea of mine). My guess is that those guys died of blood-loss.
But you can’t bleed out after the battle. Just for fun.
Often, though other forms of shock are equally possible. (And blood loss doesn’t have to be on the outside, either: it’s perfectly possible to die of blood loss without breaking the skin.)
Other good options are concussion (and resulting swelling of the brain or subdural bleeding) or various other consequences of wounds, especially broken bones. A broken rib can be a fairly… mid injury up until you twist wrong and shove the jagged end into a lung. And a broken femur is a major injury, but one of the things that makes it so major is that without a traction splint, the jagged ends are very likely to nick the femoral artery and Bob’s your bereaved uncle.
It’s also possible that they did receive a “lethal wound” but just kept going for a while: adrenaline can do amazing things.
And this is before A.I. has gotten around to designing armors for video games.
Wonderful article, but it would be improved greatly by embedded images of cuir bouilli and buff leather jackets as you did for the other armor types. I found myself flipping back and forth from google images on both of those and hoping that I had landed on “correct” examples.
D&D used to have rules where certain weapons were better or worse against certain armors but they were clunky and no one used them.
Yeah both Palladium and Iron Crown started off with add-on books to model the effects of different weapons against different armors, then branched out into their own RPGs.
I want to say one word to you. Just one word: GURPS!
(I’m well aware of the system’s flaws; I wrote the initial text file for the current-edition character builder, so I know GURPS, even if I’ve only played it once in a crappy WoT game at DragonCon. Still…it’s a simulationist’s dream.)
Thanks–this is really interesting.
I think you can’t really fix the concept of “AC” without also fixing the abstraction of “hit points,” which as structured just don’t make sense–there’s no (well, very little) inherent loss of abilities corresponding to being heavily wounded; as characters gain levels, the HP pools go up entirely out of proportion with actual changes in human survivability; they just don’t seem to track very well anything about the actual experience of being injured. You get the Black Knight threatening to bite Arthur to death.
Instead, I’ve started thinking more about hit points as more akin to “luck points,” where ‘getting hit’ doesn’t represent taking physical damage, so much as having a close call that didn’t actually connect but should’ve killed you if it had. And HP damage is thus “spending down your pool of luck” rather than “grinding your actual fleshy meat-sack.” This also explains the role of cleric-as-healer–divine magic isn’t physically stitching up your wounds; rather, it’s interceding with the gods to grant you more of your luck back; restorative potions aren’t regrowing lost fingers, they’re more like an energy drink that gives you enough restored vigor to continue barely-dodging or shrugging off future shoulda-been-fatal hits. In this context, a higher AC makes sense because better armor determines what qualifies as a close call at all, rather than reducing damage.
Now, obviously this is not the abstraction as intended, but it’s at least what I’ve come to terms with…!
> You get the Black Knight threatening to bite Arthur to death.
The Black Knight *thinks* he has D&D hit points, but he doesn’t; he actually did need those limbs to continue fighting.
As noted in comments above that basically *is* what hit points are.
In regards to the hp thing and human survivability one option could be to just say that someone with high enough is just superhuman in comparison to irl humans. There’s actually a great article(focusing on 3.5 admittigly) that makes a pretty strong case that anyone level 6 or above is more skilled(and in the case of hp totals more tough) then any real life human ever
(Admittedly this dosent apply to losing capability as you get more injured which is just part of the whole hp system)
https://thealexandrian.net/wordpress/587/roleplaying-games/dd-calibrating-your-expectations-2
“where ‘getting hit’ doesn’t represent taking physical damage, so much as having a close call that didn’t actually connect but should’ve killed you if it had. ”
Think of that guy from Saving Private Ryan who gets shot in the head, but his helmet saved him as a form of HP. But the second shot killed him.
There are a lot of systems that do something like “HP is when you’re actually in fighting shape, if HP goes to 0, or you take some kind of Critical Hit, is when you actually are *wounded*” In that sense representing stamina/vigour/luck, etc.
Statistician-and-DND-5e-GM here (bow before the epitome of nerdery!).
Sorry, but in DnD5e one point of AC does *not* reduce your chance of being hit by 5%. How useful it is depends on the opponent’s attack bonus.
Consider a naked character with an AC of 10, and suppose they raise it to 11. Pit this character against a weak opponent with a +0 attack bonus. At AC 10, the enemy has to roll 10 or more to hit on a twenty-sided die (a d20); at AC 11, the enemy has to roll 11 or more to hit. The chance goes from 50% to 45%, a 5% percentage *point* improvement, but a *relative* improvement of 10% – out of ten attacks that would have hit at AC 10, one will *not* hit at AC 11.
However, our character has survived this encounter and now meets a foe with an attack bonus of +10. This new opponent will add this bonus of +10 to their d20 roll, so the lowest they can roll with the modifier is 11. They will *always* hit, regardless of the roll, except that a rolled 1 is an automatic miss. Whether our character has an AC of 10 or 11 makes zero difference whatsoever to their chance of being hit by this opponent.
Conversely, suppose our character is much better armored and has an AC of 20. They come across a magic ring that would increase the AC by one point to 21. Should they expend one of their three attunement slots to attune to it? Well, if they next meet one of the weak foes described above, with an attack bonus of +0, this ring will again make no difference whatsoever – at both AC 20 and 21, the opponent will only hit if they roll a natural 20, which is an automatic hit in DnD 5e. However, it *will* of course make a difference against that stronger opponent with a +10 attack bonus.
Sorry about adding that bit of unmitigated pedantry to the discussion… now back to reading the rest of your presumably excellent essay (did I mention I am a Bayesian and by now have very strong priors about the likely quality of what I am going to read here?)…
(Incidentally, I read a paper by Taylor 2023 in *Libyan Studies* today, which I believe I found recommended on your blog. Thank you very much for such little recommendations, I enjoyed that article immensely; the description of the different incentives of Carthaginian and Roman generals to engage in battle very much reminded me of the incentive structures in place in the British Royal Navy in the age of sail, and the fate of so many unsuccessful Carthaginian generals could be compared in interesting ways to that of Admiral Byng of “pour encourager les autres” fame…)
That article on Carthage vs Rome was indeed fantastic.
This blog needs more such unmitigated pedantry in the comments.
Now, I agree completely, but I think it’s more interesting to include some examples which are not literal edge cases to make the point.
So AC 15 vs AC 16 vs the same +0 attacker.
AC 15 gets hit on [15, 16, 17, 18, 19, 20]
AC 16 gets hit on [16, 17, 18, 19, 20]
That’s 6:5 improvement.
AC 18 vs AC 19
AC 18 gets hit on [18,19,20]
AC 19 gets hit on [19,20]
That’s 3:2 (same as 6:4) improvement, much better. So the more AC you have, (and the worse your attacker), the more the marginal point of AC matters.
This is actually one of my problems with the D&D armour system. To an average person, it might seem like a +1 to AC should give a similar increase in survivability to any character, but in reality, each point by which you increase your AC gives a greater increase to survivability than the last, until you hit the point where your enemy has to roll a 20 to hit you, at which point it stops mattering entirely.
This is why I think that bounded accuracy is one of 5e’s best ideas (although it introduces other problems). Because the hit chances tend to be within a limited range, a +1 to AC does give a similar increase to survivability to any character, and it’s very unlikely that you’ll hit the point where your hit chance is either 5% or 95% (and bonuses stop mattering). This was not the case in 3.5e.
Now, if I were remaking the system from the ground up, instrad of using the d20 system, I might have something like, roll 4d6, for every 6 you rolled, roll another die and add it to the total (and if it’s a 6, roll another one, and so on). This would mean that there’s an exponentially decaying “tail” of probability that extends to infinity, so you’ll never reach the point where a +1 to AC stops mattering.
Are there historical examples of armor that did have a second, lower front plate, even it started lower down than it does in the Baulder’s Gate artwork? That was my understanding, though possibly that’s it’s own myth perhaps resulting in some confusion over what a fauld is and/or where the fauld stops and the tassets begin, in a fauld-and-tassets setup.
The Dendra Panoply did (it actually had three lower plates), although in that case the plates at least overlapped (there wasn’t a ridiculous gap where someone could stab you in the gut). Also, the Dendra Panoply was a peculiar design with little influence on later armors, and which I highly doubt was influential on this D&D artwork.
I think computer games shouldn’t even attempt to replicate tabletop systems. After all, for the sake of playability tabletop systems have the ideal of resolving hits with a single roll and looking up static values from tables, while computers can handle arbitrary amounts of bookkeeping and rolling. In contrast, tabletop systems can rely on a human DM to simply arbitrate trivial cases of resolving actions, while computer games can’t do that and have to at least fail gracefully. Good simulation of plausible outcomes is also a strong desideratum to facilitate emergent narratives: if not accomplished, any roleplay has to work against, not with, the system. (For the record, systems INSPIRED by tabletop can be good: dice for example have several desirable properties, such as naturally working with low values that can be mapped to concrete real-world meaning like e.g. plate armor has 20 protection, allowing the addition of numerous small but meaningful modifiers with little ambiguity about stacking rules, and probability-distributions that are easy to comprehend thanks to our familiarity with dice)
For example, as the author notes, a big problem with the AC system is how it functions by entirely negating the damage, being incapable of mechanically distinguishing between hitting a fly and a sleeping dragon (3.5e additions of flatfooted AC and touch AC gesture towards something to this effect, but don’t do a very good job at it): a problem that arises from the desire to resolve hits with a single roll and no further computation (but somehow tabletop D&D still manages to be pretty rules-heavy…)
As an example, fantasy wargame Dominions uses opposed rolls system with exploding dice: that’s necessarily twice as much rolling (since both the attacker and defender roll), rolling a six means you may have to roll a die more than once in any given part of hit resolution, and there’s multiple rolls involved in the damage resolution sequence. But none of this is a gameplay hindrance because the computer does the bookkeeping, it actually makes the system more intuitive (by for example MAKING that distinction between hitting a sleeping dragon and a fly, and mapping the rolls concrete real-world concepts unlike e.g. AC that can’t decide if it’s about risky combat style presenting potential openings for the opponent or worn armor, and by using the same “DRN” roll in all occasions while D&D for instance can’t even fully commit to d20), and produces rather nice probability-distributions. For example, once it has been ascertained that the arrow has struck its target, an arrow fired from a longbow has 86% chance of not doing any damage against a target wearing plate armor, 14% chance of doing at least one point of damage, 11% chance of doing at least two points of damage, 8% chance of doing at least three points of damage… and ~.5% chance of killing a human wearing plate armor outright. Naturally, these values haven’t been achieved by gathering data about the lethality of longbows and then fitting the damage resolution formula to match that data, but referring to Arrows vs Armor videos produced by a collaboration of Dr Tobias Capwell, Joe Gibbs, Tod Todeschini, et. al. that shows breastplates and helmets are effective proof, limb armor offers incomplete protection, and outlier hits may make it through e.g. the breaths in the helmet, it does reproduce the pattern of armor mostly doing its job at protecting the wearer, but at the same time not being invulnerable (unlike flat unrandomized damage reduction is prone to do depending on scaling).
And this by no means is the perfect system. But it is an example of a system that has been designed under no illusions of having to kowtow to tabletop system constraints, or to shy away from potential of a computerized system.
As a long term proponent of the idea that tabletop mechanics infesting digital roleplaying games are greatly negatively impacting the design space, I agree with most of this. Although I prefer no dice systems. But dice systems *can* work.
The real problem is that the older tabletop properties are so powerful as IPs that they create a black hole sucking in high production value dev dollars.
The advantage of tabletop systems is that they’re fairly comprehensible and it’s easy to understand what things do, while computerized systems can be difficult to follow.
Right. An example I like to use is this formula for damage calculation:
DMG_f = DMG_b × (1 + I1 + I2 + I3 + I4 + I5) × (1 – R1) × (1 – R2 – R3) × (1 – R4) × (1 – R5) × (1 – R6) × (1 – R7) × (1 – R8)
Which is from Heroes of Might and Magic 3 and in actuality it’s not as bad as it looks, but crucially, I couldn’t for the life of me recall the formula from memory, I can’t point out to you the terms that cause Tazar’s Armorer speciality to outscale Crag Hag’s offense at some high but possibly reachable level, and honestly I don’t even have a clue how much an increase in attack skill increases damage, only that it does, and mostly getting by having built up intuition over the course of having played the game for thousand hours plus. Most no-dice systems are more comprehensible, but a formula like this is always looking in the design space unless you’re careful, while die-based systems are comprehensible pretty much by default. Even if you have unintuitive monstrosities of a mechanic like THAC0.
If your system is too complicated, then you interfere with one of the dopamine centers for a video game:
“Ok, I’ve killed these guys. Time to loot!”
“Is this weapon better than what I currently use? How about this armor piece?”
If it’s based on a tabletop system, it has to be simple enough for someone to grasp fairly quickly.
A computerized system could be much more accurate, but it risks becoming so complicated to figure out that changing weapons becomes like going through the US Army rifle procurement system.
It doesn’t have to be that complicated, but then you’re either using a tabletop system, or designing one that could be used as one.
I played a little in the Pathfinder Online game, where they did throw out the tabletop system and used their own thing (while naming things like in the tabletop system, just to increase confusion). You may be able to find descriptions online, but questions like “How much better is a +1 sword” ended up being “It’s better. How much? Don’t know”
Typo:
“might attack to”
“might attach to”
Way way back, one of the first supplements to pamphlet sized D&D did have a table of bonuses and penalties for different types of weapons attacking different kinds of armor, but my impression is everyone thought that was too much detail and ignored it, leaving it in the dustbin of gaming history.
I always assumed that ‘banded mail’ was lorica segmentata. There are 5e versions that incorporate damage reduction, like Dragon Heresy.
I do note that even BG3 does so: At higher quality levels armour will start to attach some level of damage reduction.
The concept of AC instead of Damage Reduction seemed very exotic to me when I first played D&D. I cut my teeth on the Swedish RPG Drakar och Demoner, where the rule was damage reduction.
Bravo.
Hey I’ve got a really helpful suggestion: why don’t you follow up this long article analysing the historical accuracy of armour in a video game by writing another one about a different video game? I’m of course thinking of Kingdom Come: Deliverance, of which from your previous mentions I gather you weren’t a huge fan from a gameplay perspective (it certainly has its flaws) but which as I recall has pretty much the right system? No doubt an expert in 15th-century Bohemia could point to all the glaring inaccuracies I missed but they did at least seem to put an awful lot more work than any other RPG into getting the history right; I’d certainly never seen the word ‘brigandine’ before I played it. Of course it’s not your area but if I were to see a guest post on the topic in my inbox I’d be clearing my schedule.
On a totally different note, this is my first comment but I’ve been a longtime lurker. Your series on Minas Tirith and Helm’s Deep inspired me to reread LOTR for the first time since childhood, to my great benefit. I enjoyed rediscovering Tolkien so much that when I get to the bottom of the pile of books I’m forcing myself to read before I buy another (a common humblebrag but I’ve stuck to it for over a year, and the pile is <10 now) the first book I'm buying will be the Silmarillion. I find your blog consistently fascinating, and I say this as someone for whom the acquaintance with academics that his bachelor's degree provided was enough to run screaming out of academia with no intent to ever return. So – thankyou.
I want to see a critique of KC:D but by a different historian than me, focused on assumptions about social structures and ethnicities rather than on arms and armor, to be frank. KC:D’s armor is mostly spot on – most of it are straight-up copies of museum pieces (some a bit anachronistic), so there’s not a lot to say there. I think it is the deeper assumptions about the shape and nature of medieval society in KC:D that are more worth discussion, but again, probably fall to a different sort of historian than me.
Perhaps a guest post one day.
>Brigandine/Metal Lamellar
I’m glad to see these grouped together. I’ve noticed that western (non-academic) discourse on the subject tends to prize brigandine as somehow different and better than other forms of lamellar armor, which reads as little more than residual European chauvinism, which is pretty rife in these discussion spaces.
As the owner of a brigandine myself, I have to say that it’s far more durable and sturdier than lamellar armour. Lamellar’s durability in combat is heavily constrained by its leather/textile laces, in a way that brigandine is not.
As a side note, I’d like to add that brigandines were widely used in Late Imperial Ming and Qing China.
As the owner of a brigandine myself, I can say that it’s far sturdier and protective than lamellar armour. Lamellar’s durability is heavily limited by the use of leather/textile laces compared to brigandine’s rivets, while the latter’s larger plates protect better since they allow the blow’s strength to be distributed on a larger surface.
I’ve never met this “residual European chauvinism” in any kind discussion whatsoever on the “lamellar vs brigandine” topic (I’m Italian myself), but it’s worth to note that brigandine was the standard armour of choice in Late Imperial China, in the Ming and Qing dynasties.
As a proud owner of a set of lamellar armour: Brigandine often (but not always) has bigger plates than lamellar armour. Bigger plates will always offer more protection than smaller ones, as they are able to distribute the force over more area.
You may have enjoyed looking at 2nd edition and 1st edition D&D armors and their interactions with weapon types. The overall armor system still had the “AC” aspect (which is a strange combination of ‘didn’t hit’ and ‘didn’t do damage’, I’ll support) but there were also different ways that, say, Bludgeoning Damage worked better against heavier armors and the like.
There’s also (at least in 3.5) optional rules for ‘Armor as Damage Reduction’ that I’ve always wanted to play with, since, as you say, that makes a lot more sense. It’s what I’m incorporating into my own system, even.
Aaargh, flashbacks. Yes, AD&D had the full page weapon vs armour table in the back of the DM’s Guide, which would tell you which of a glaive or glaive-guisarme got the extra +1 vs splinted mail.
A nice reasonably well thought table that almost nobody actually used but it really served to break up the current ‘sword is good vs everything’ meta.
And there was also the wonder of how it went from 10 for no armor down — to zero, and past it.
Inspired by a naval war game, where Class 1 was the best armour, and Class 10 was the worst.
Naval war games didn’t have to worry about magical bonuses.
Someone suggested a neat shield hack for descending Armor Class (aka THAC0): a shield divides your AC by 2. So if you’re unarmored, AC 9 or 10, shield is very useful. If you have mail, AC 5, it’s somewhat helpful. If you have full plate, AC 1 or 0, it’s time to graduate to two-handed weapons.
One more thing!
I have always been SO ANNOYED with the ‘flexible armor’ in videogames. Ironically, as graphical fidelity and animation quality have improved in games, the visual quality of heavy armors has REDUCED- breastplates and plate armor, instead of being solid, will FLEX with the torso! It’s horrible!
A couple more years of GPU speedup should take care of it.
Simulating/calculating the movement of bodies, clothing, and armour takes a lot of computing power. Easiest case is rigid body animation, basically robots, which is also a good match for people in full powered armour (Halo) or full plate armour. Bending is restricted to well defined joints with limited movement.
Soft body animation, eg people not in full plate armour, is harder because body parts don’t just rotate around joints, they also change shape as they do. (Look at your elbow as you bend your arm.)
Clothing and hair adds more problems, and much much easier if everyone has a buzzcut hairdo and tight fitting clothing which just flexes with the body. Loose clothing, which also includes loosely fitting armour, complicates everything because you now have bits that bend and bits that don’t, and the armour/clothing might limit the movement of a limb, might not.
Right now our PCs and consoles can do rigid body animation, or soft body animation with some clothing/hair, but not realistically combine the two.
I also suspect that since a lot of computer armour is less than entirely practical, any realistic simulation would show sharp metal edges digging into sensitive bits of anatomy. Not what gamers are really asking for.
I notice that none of these depictions include a helmet. Does the game just add helmets independently of armor?
In BG3, helmets are a separate armor piece that usually adds a bonus to something, but not to armor class. They can also be toggled on/off for cutscenes and for general gameplay, which is a nice quality of life feature: the character can wear a helmet in combat, but take it off in conversations. Hand and foot protection are also separate accessories and not part of the armor, and also do not usually add to AC.
Yes. Helmets are separate items and may have magical properties (or not). They are equipped separately from armor, as our gloves and boots.
Good guys in high fantasy have invulnerable heads and therefore do not need helmets.
They’re separate items, yeah. They appear on models by default, but you can toggle settings to hide them.
Penny for your thoughts on Pillars‘ flat damage reduction armor system? Sounds somewhat like what you’re talking about at the end, but without the conversions between damage types.
I agree that “Studded Leather” should have been dropped from the game decades ago.
I recall that there was a version of quilted doublet that had small squares of metal sewn into it (not as extensively as Jack of Plate, but less dense with the number of metal squares, 1cm x 1cm ish squares, maybe 20-25% of the surface area of the doublet, mainly at intersections of the quilting pattern). I saw one in person, but I don’t know if it was historical or not. It wouldn’t be very useful against stabbing, thrusts, and arrows, but it seemed like it would be somewhat useful against slashes. For simplicity, I’ll just call it a reinforced doublet.
It has long been my thought that Splint is supposed to be Brigandine and/or Coat of Plates. Wikipedia does say Splint was a thing (and somewhat similar to Brigandine).
In D&D 5e, Scale is supposed to include Lamellar (D&D 5e is painted in broad strokes).
As for Ring armor… I know I have seen something like it in shows (like “Vikings”, where Ragnar wears an armor that looks like semi-rigid leather with metals rings secured to it, in leather braids along its surface), but that doesn’t mean it was ever real. It could just be a case of tv writers imitating D&D. Maybe that should be replaced by Jack of Plate, as a Medium armor.
There’s also some amount of game design in the D&D armors. Why wear Half Plate if it’s going to have most of the drawbacks of Plate (stealth disadvantage, no dexterity bonus, more expensive than anything other than Plate)? Part of the value of the Medium Armor category is that you get at least some of your Dexterity bonus to Armor Class, there’s no Strength requirement, and you’re less likely to have a stealth penalty. So the incentive of Half Plate is: while it’s not as protective, but at least you’re in the Medium category. That split goes through most of the set of Medium vs Heavy armor: a lighter version of the Heavy armor is in the Medium category, giving you the trade off of better AC with Heavy armor costs not just in gold, but in stealth/dex/str.
I would like to see Leather in each of the 3 categories. Buff in the Light category, Hardened Leather Lemellar in the Medium category, and Hardened Leather Plates in the Heavy category. Or maybe go the other direction: Hide (layered pelts) is the light “animal skin” armor with an 11 AC. Buff as the low end of Medium Armor made from Leather. Hardened/Boiled Leather Plate as the low end of Heavy Armor made from Leather. Then don’t differentiate what the scales/lamellae are made of: it’s all just what D&D calls “Scale”.
And I would love to see the descriptions of the backings and accessories fit history better.
Maybe this:
Light (all get the full dex benefit):
Padded/Gambeson – AC 11 (can be worn under medium and heavy armors)
Lightly Layered Hide – AC 11
Buff Jerkin – AC 12
Medium (all have max dex benefit of +2):
Heavy Layered Hide – AC 12
Buff Coat – AC 13 (can be worn under metal armors)
Mail Shirt – AC 13 (stealth disad)
Scale/Lamellar Coat – AC 14 (stealth disad)
Cuirass/Jack of Plate – AC 14
Half Plate (Steel) and/or Boiled Leather Plate – AC 15 (stealth disad)
Heavy (all have str minimum, no dex benefit to AC, and stealth disadvantage):
Mail Hauberk – AC 16
Brigandine/Coat of Plate – AC 17
Plate – AC 18
The note about wearing it under medium or heavy armor is that you can wear them during downtime since they’re comfortable … but if you’re wearing gambeson already it should make the process of donning your regular medium or heavy armor faster than if you were wearing plain clothes. But the small metal squares of a Reinforced Doublet wouldn’t really be suitable to wear under an outer armor (in my estimation).
So was Tolkien talking nonsense in “Farmer Giles of Ham”?
Where would the nonsense part be in that?
“If it’s real ring-mail you mean, then you can’t have it. It needs the skill of the dwarfs, with every little ring fitting into four others and all. Even if I had the craft, I should be working for weeks.”
This is an excellent description of chainmail. The other idea, in its application to their personal situation, is explicitly an improvisation in-story and strikes me as a plausible compromise. The claim that this is a routine method elsewhere is more iffy (surely scale or brigandine would be a better choice?), but it is hearsay in the context of the story, or perhaps the smith is knowingly lying to establish a fake tradition behind the ad hoc solution.
Tolkien wrote Farmer Giles of Ham in 1937, when the armour specialists he would have been able to consult were sure that ring mail was a thing. (Not published until 1949, but the intervening years are not exactly a boom time for medieval archeology and history.)
He did the best he could with the knowledge available at the time, which is all any of us can do.
Nonsense, I see endless people online excoriating authors for not predicting future changes in theory.
(I do wonder why they don’t just list the two or three perfect books and leave it at that.)
I do note that IIRC; Gygax was largely working with older library and that were probably not that dissimilar from what Tolkien used. (I seem to remember a thread on RPG.net that specifically tried to figure out which exact books he used for reference, and where the problem was his reference, and where the problem was his interpretation of his references)
Tolkien here may be drawing on Norse stories and poems, where ‘rings’ can be used for chain armour.
Yeah. This sounds like Tolkien was just using “ring-mail” as a way of saying “mail, and it’s made out of rings,” in the same sense that a “battering ram” is a ram for battering things. If someone says “the raiders picked up a ram and knocked the door down,” it’s clear from context what they’re talking about. But if the same person says “the raiders picked up a battering ram,” it doesn’t mean they mean to say “there is a special kind of ram called a battering ram” as a way of making up something that isn’t real in the source material.
Tolkien was generally pretty tuned in to the available information at the time and his worldbuilding and attention to detail could get downright obsessive.
But he “talked funny” by modern standards. First, because he learned the English language around 1900. That is, to be clear, now roughly 120 years in the past, roughly 30% of the distance between us and Shakespeare. It is no surprise that vocabulary and usage have shifted with time. Second, because even by the standards of the time, Tolkien was often influenced by archaic sources, and by this I don’t mean “badly informed,” I mean “old in style.”
If there’s a plausible explanation for how Tolkien could know what he was talking about and just use phrasing foreign to us as modern readers, I tend to accept it.
That doesn’t even get into how armor has varied wildly across editions, with some armor types (chainmail and scalemail in particular) varying their armor capability back and forth each time.
On the leather uncomfortable thing, something I’ve seen noted was Native Americans tended to drop buckskin for textiles, particularly for shirts. And that is pretty nice leather too! I wonder how much essentially larping was involved in a lot of frontiersmen going for buckskin. At least some of it was from what I remember.
Oh, boy, Edition Wars.
(I include D&D vs Storyteller in that along with D&D 4 vs D&D 5 and the like)
I think I’m the only person I know who likes Armor-as-reduced-chance of being hit. 😛
Mostly because I mostly play 3.5/PF, and honestly damage reduction just doesn’t scale enough to matter. When hits regularly do 40-120 damage per shot, damage reduction in the 5-15 range doesn’t actually do anything to slow down how fast you go down.
Then again, AC also kinda ceases to matter past the mid levels in 3.5/PF, because to-hit bonuses tend to scale faster than AC bonuses. Which leads to the meta being, much like Pokemon Gen 3, Rocket Tag – optimize for high speed/initiative and overwhelming offense. “Go second and die.”
In this sense D&D 4 and 5 are a big improvement, because they don’t make glass cannon the only viable solution in the optimization meta. But I just don’t enjoy bounded accuracy. So I mostly play 3e derivatives with groups where we have an informal agreement on where we want to set the Killocity and no one is allowed to diverge too far up or down (DM included).
Do they have male knights in this game? I know warrior women are hot but a slip of a girl in clingy armour is not a convincing paladin.
As a CRPG in the western tradition, Baldur’s Gate III makes your Player Character fully customizable. So the PC can be a male knight, certainly.
Also “Clingy Armor” really?
BG3 is a huge improvement in terms or armor-that-actually-covers-things compared to 99% of MMOs. That’s really a mis-aimed complaint here.
Your comment reads more as misogyny than considered criticism.
Also, as it seems to be directed at Dr. Devereaux rather than myself, I don’t think it was meant to be a reply to my comment post?
Can a woman be misogynistic? A slip of a girl isn’t a convincing knight, a tall and husky woman would be more convincing. I’m sick if the waif look in fighters.
Eh, the increase in ability of all fighters show that it’s heavily magic. Magically empowered waifs work.
Ip Man was 5’4 and pretty skinny – at 120lb, borderline underweight in BMI terms. I think he made quite a convincing fighter, and his student Bruce Lee would agree.
The character illustrated looks quite a bit taller and heavier than that (difficult to judge from a computer graphic).
Yes, women can be misogynistic.
(The patriarchy really wouldn’t be sustainable if only men perpetuated it.)
If you didn’t mean to be, well, tone and intent are hard to convey and interpret on the internet.
And again, BG3 is already much less waif-ish than most computer games in the RPG genre. Certainly there is much room for improvement, but, well, credit where it is due.
In any case, it’s a D&D setting, so physical musculature and actual applicable strength are not remotely correlated.
It’s a world where Pegasi fly using their wings and this is considered an (Ex)traordinary ability, not a (Su)pernatural one, and works just fine in anti-magic fields.
PS, I thought I was replying to Bret. Sorry.
I have to ask – are the problems highlighted above equally apparent in the male character models, or are any of them specific to the female designs?
As we no doubt all know, video games and fantasy art in general have a habit of “sexing up” the female designs more than the male ones, and even where this isn’t just about showing more skin it often manifests as designs that better show off the character’s figure. And of course, with male armours there are a lot of historical designs available to straight-up copy, whereas (while in fact it usually probably looked almost identical) female armour designs invite more imagination.
I seem to recall that the armour designs for the male characters in the original BG1 were generally pretty ok (once the studded leather, etc. were accounted for), but the female models’ armours were much more figure-hugging. The BG2 models were just awful generally.
I did notice that the Githyanki half-plate loses pants if a woman wears it.
While I haven’t played Baldur’s Gate 3, in Dungeons and Dragons Online the armour for females is often skimpier than for males. There are some good armour appearances, but plenty of silly designs.
The high slit is almost certainly for animation’s sake. That might indeed also why they go with the weird two-plates thing: So characters can bend down and pick stuff up without looking really weird or having to make two separate sets of animation for different types of armour.
I seem to remember there being a thread on RPG.net about some of Gygax’s sources (specifically polearms?) and some of it was “His sources were old books based on victorian stuff” but also some direct stuff that he just read wrong.
Incidentally I’ve always thought a byrnie meant a slightly larger garment than just a short: Something that extended at around knee-lengths.
Another point is that D&D and similar settings aren’t neccessarily *military* in the same way, and you should probably also be looking at civilian clothing and such. I know WHFRP uses “Soft Leather” (bad name) to basically mean “protective/heavy clothing” IE: not real armour but something that might still protect you from cuts or scratches, (Given that Warhammer characters are more likely to be rat catchers, bonepickers and lawyers rather tahn soldiers this makes sense)
Also, i think it’s good to remember that the AC system didn’t originate for role-playing games, but rather for mass-scale combat: The abstraction makes a bit more sense there.
(and it’s not the only abstraction, HP, attacks, etc. are all to some degree abstarctions, which gets espeically noticeable when you go back to the old round/turn distinction, where one turn is 10 rounds)
Irrelevant aside: I keep misreading the title as “The Gap Into Armor” and think it is about another Stephen Donaldson science fiction novel.
Ah, the Gap series. A series I think is, on some level, quite good but because of the subject matter I also *really* can’t recommend to anyone.
But then again, that goes for most Stephen Donaldson stuff.
Back in the day Rolemaster armour was reasonably realistic: Heavier armours made you easier to hit, but harder to damage significantly. Compared to unarmoured, a heavy armour user would take damage a lot easier, but was much harder to crit (and crits were what won most fights).
Since then I’ve just decided to cope with the handwavy way basically all other d20 systems treat AC and damage.
Taking a moment to mention Warhammer Fantasy Roleplaying system, which both had armor-as-damage-reduction _and_ layered armor! “Full plate” didn’t just involve rigid metal plates, it also involved a full set of chainmail armor and a base layer of… leather armor.
Look, it’s not a perfect system. But it does layer!
Intriguing why the game is written out here as “DnD” when “D&D” (ampersand) has been standard for decades? Is this some editorial faux pas or of deeper significance?
I post this as an amicus, who was also once a game design staffer at TSR Hobbies, when that Wisconsin publisher was the original purveyor of Dungeons & Dragons.
I’ve noticed DnD/DND rising in popularity recently, probably because it’s easier to type on a mobile keyboard.
It is surely not the inspiration for d&d, but the Tlingit made a type of armor from leather and metal coins that resembles something like a half way point between studded leather and scale armor.
Beat me to it. It wasn’t Tlingit I don’t think, but rather east Asian, using Chinese coins. Anyway, the original AD&D drew quite heavily off the old Stone’s Glossary. Under armor, it clearly shows leather coats with coins or other metal sewn on. I agree that brigandine should be in, but metaled-up leather armor was and is a thing (nowadays, for punks and some motorcycle armor, where the studs hit the pavement before the leather does, to save skin in accidents).
Oops, didn’t read far enough. We’re both talking about “ring armor”, with the coins being rings. I do think talking about D&D without Stone’s Glossary to hand is missing quite a lot. Anyway, yes, ring armor technically existed, but not in Europe. You could make it now by sewing washers onto a shirt.
First of 3 videos on Iroquois wooden slat armor https://www.youtube.com/watch?v=qmkQS8_MODE
lamellar-ish
Excellent article. I was surprised by the thrice-repeated claim that D&D players would significantly start viewing armour in terms of how it is represented by Baldur’s Gate. I haven’t heard much interest in the game at all from my various TTRPG groups and online contacts, and I’m certain a strict majority of them, not being video gamers, haven’t heard of it.
My impression for this kind of thing is that tabletop gamers aren’t actually more interested in RPG video games than they are in other genres. The people I know who play both TTRPGs and video games mostly enjoy puzzle games and co-op shooters.
Nice take-down of these ridiculous fantasy armors, that give rise to so many unpaintable miniature figurines and nonsensical computer images. People who buy into this are being duped and misinformed. But I suppose Westerns on TV did the same thing for the Wild West, eh?
PS: I think the baleful influence of Renaissance Faires and cosplay and CGI effects and “Goth” fashions have played a part in this as well.
If we are being realistic, heavier armour does not make you easier to hit (but harder to damage). Dodging in the midst of battle was near impossible, hence everyone armoured up as much as they could afford. A good many famous warriors failed to dodge an incoming spear or arrow (looking at Harald and Styrbjorn).
I do note that D&D does not usually represent that kind of mass combat though: We’re not talking thousnads of men in close formation but rather small groups of people having a brawl in a back street.
D&D is derived from some man-to-man hacks to group wargames like Chainmail or Strategos N, where you roll to see how many men are hit (and thus disabled/dead), and in which better-armored units requiring more simultaneous hits to whittle down. The big thing that most people miss is that what the game calls a “hit” is always a death-blow (or at least a disabling blow). A non-hit is not necessarily a miss, but rather could be a blocked, deflected, absorbed, or minimal-impact hit.
On the personal level, D&D breaks that down into “am I hit?” (which is decided by the attack roll vs. AC) and “how many hits?” (which is where N hits are replaced by N hit dice, or Nd6 hit points). Here, a “hit” has gone from being sure death to being something that has a fair chance at death. (Initially, this was while you subtract d6 damage from your pool of d6 hit points.)
The attack-vs-AC-hence-damage paradigm assumes that the armor is doing its job and blocking/absorbing the blow most of the time. And that’s valid for the way medieval armor worked: if your breastplate wasn’t made of papier-mâché, then either the arrow hit it and was defelected, or hit somewhere else and injured you. There was no real risk of the arrow hitting your breastplate and yet still plunging through it into your chest. That’s the sort of thing you see when you get relatively high-powered firearms.
Re chain shirts and chain mail: I thought the difference (in D&D 5E that is) was supposed to be that a ‘chain shirt’ is, well, just a ‘shirt’ – that is, a mail haubergeon or hauberk – while full ‘chain mail’ was full-body mail including long sleeves and mail leggings? Classing the former as medium and the latter and heavy makes sense to me.
I feel like a chunk of the misunderstanding here can be chalked up to not realising that hardened leather is a lot closer to metal both in its level of protection and its flexibility and weight than most people realise. That explains the underuse of leather in the heavier classes, and the ‘leather is badass’ explains the overuse of leather in the base layer and the lighter classes.
Yeah, I’m definitely feeling a “chronic lack of lower body protection” from a lot of these armors. Some of them fail to cover places you really, really don’t want a spear going unimpeded.
A topic that might be cool to address in future is to what extent armor (and weapons, etc.) would have been different for “adventurers” historically versus soldiers. To the extent that there were such people or that we have evidence for them, anyway. If someone was doing robbery or assassination or something, would they have bothered to wear armor and if so what kind? You talk about which armors are more or less physically constraining, but you kind of dodge the question of which armors are better or worse for stealth, despite that being pretty central to how 5e distinguishes them.
The central question to break down is how your adventurers avoid getting slapped around by the general group of people they try to prey upon.
– They are (part of) an army. Probably not professionals, but the equivalent of fyrdmen.
– The opposing army was already defeated, they are an occupation and/or police force.
– They are raiders with better operational mobility than their target, i.e. while they can overwhelm a small town’s militia, if the actual army does show up then the plan is to run away. VIKINGS! Pirates! Horsemen!
– Somehow the legal system makes a particular kind of crime practically unpunishable, e.g. shepherds “employed” by the elites on long-term tallies can rustle each other’s sheep back and forth.
– They avoid being identifiable by their victims, and it would decrease tax revenues more to kill everyone in a day’s walk than to suffer their crime. Hide not among shadows, but among law-abiding people. Classical banditry.
– They hide not their presence or identity but their intent to kill the local tyrant, and expect that (unless he has guards and, success or failure, the guards kill them before they can get away) the tyrant’s faction will fall from power after the assassination.
Massively over-simplifying, people who aren’t active duty soldiers won’t have armour and are carrying a sword or quarterstaff at best. The local authorities, whether a manor, small town, or city; aristocratic, oligarchic, or democratic; do not want heavily armed people who aren’t under their command wandering around the place!
Wearing armour signifies that you are looking for a fight. Today, even in a city with very permissive gun ownership laws, wearing kevlar body armour will attract attention from local law enforcement. Walking into a bank wearing a motorcycle helmet is not a good idea either. Medieval authorities had similar attitudes. Plus, armour is heavy, hot, uncomfortable: wearing a suit of armour all day is not pleasant.
(And yes you can hide some armour under civilian clothing, and a quilted gambeson type armour makes a good winter jacket. Massively over-simplifying.)
Weapons are likewise restricted in urban areas, although there’s variation over time. Bret gave the example a couple of posts ago of bodyguards in the city of Rome only being allowed to carry clubs. Medieval cities might ban swords, might allow nobles to wear swords. Definitely not going to allow maces or polearms or bows. (And again, those are heavy and awkward, why would you want to carry one around all day?) Commoners can carry a stick, although probably better not to have obvious metal reinforcements.
Since no-one has armour, a sword or quarterstaff is all you need. Long stabbing swords are very effective, hence rapiers. I believe English town records also show a large number of people being killed with quarterstaffs.
So amusing thought in relation to this, thanks to D&D characters having cantrips and other abilities (ex.unarmoed defense) some d and d charecters might be comparatively heavily armored just by their inante abilities or training unlike in people in reality. You can tell the fighter to put away the armor less so with the barbarian who’s skin is so tough it acts as armor, or say the wizard who has combat cantrips memorized and doesn’t need a spellbook to cast them basically at will
In a world where magic is real, the city guard routinely deals with these issues.
– They issue debuff and antimagic wands, as the fantasy equivalent of tear gas. Civilians may also carry some.
– They mandate that adventurers of this kind wear “cursed” debuff artefacts (e.g. medallions) while in the city.
– Failing that, they deny entry to such adventurers.
On the other hand, there were eras where men were definitely expected to carry swords in some times and places. A woman could just go into Versailles, but a man had to carry a sword. Hence, there was a royal official who made money in his rather brisk business of renting swords.
Broadly true. Bodyguards (of nobles and merchants) often carried pole-arms; in some places anyone with claim to nobility status wore a sword as a mark of status (hence the noted image of the Polish poor noble tilling his field while wearing a wooden sword), people commonly wore a blade rather like a cutlass (a ‘hanger’) – for defence while travelling, axes and hedging bills were everyday tools and spears used for hunting boar and wolf. So there was a fair bit of weaponry around, and often the skill to use it.
That’s not true for most of medieval europe. People above a certain level of wealth were expected (sometimes required) to own a certain standard of arms and armor so that when called to fight, they wouldn’t be useless. They wouldn’t go around wearing armor, because it’s heavy and uncomfortable, but they’d have a family suit of mail at home
That’s what I wrote. In the context of adventuring and what to expect on the random encounter table, people are not “active duty soldiers”.
An rpg blog I enjoy has a very good post collating different types of armour, worth a look if only for the pretty images: http://middenmurk.blogspot.com/2015/09/body-armour.html
That’s a fascinating page, thanks for sharing! Everyone should take a look!
There are examples of what seems to be studded armor from India (although variants of the chilta hazar masha also incorporate splinted plates), and what seems to be ring armor from Sudan. The Tlingit scale armor made from Chinese coins is really cool. And the Brewster Body Armor is aesthetically hideous but could apparently withstand fire from a Lewis light machine gun.
You might enjoy Fantasycraft (a fork of 3.5), which treats armors as entirely for DR. It also cuts out trap feats, while still keeping complexity & allowing you to do fun things (and gives the gm tons of leeway on how realistic, fantastical, grim and/or hopeful of a story they want to tell).
The thing that bothers me about the “armor as DR” approach is that it assumes that the attacker is making the same sort of attack against an armored foe that they would against an unarmored one. Yes, if you slash across your opponent’s stomach with a sword, their armor will turn that into some fairly minimal blunt trauma — that’s why you don’t do that! Whereas, if you attempt to strike at a gap in the armor, or use an anti-armor technique such as the murder stroke, you are faced with increased challenge in successfully landing a telling blow. Thus, increased difficulty to hit.
This is modeled in crunchier DR-based systems as called shots, usually, but D&D is not that crunchy. For the level of detail that it tries to simulate, armor-as-AC actually seems more realistic if we assume competence on the character’s parts. (What happens when you do successfully land a blow is… not. Hit points are a terrible system.)
+1 to this. I think the default D&D system is fine and prefer it over “armor is damage soak.” If you’re going at a dude in heavy armor with a dagger / stiletto, you’re not trying to bash through his armor, and if you try you’ll be dead. You’re trying to slash any open weakness you can find – and the fuller the armor, the harder it is to find such a weakness. Miss = you didn’t find a weakness, hit = you did. If you did “armor as damage reduction” than low damage daggers just bounce off for no damage forever, which is unrealistic – you can still get a lucky shot in at a vulnerable place.
That said, I do agree that the D&D style doesn’t work for certain weapons, most notably maces which really are designed to smash into the armor but turn the fleshy body into goo anyway. Maybe stuff like two-handed greatswords, too. As others have noted, 2e D&D had a special per-weapon table that could hypothetically give maces bonuses to hit vs. heavy armor and the like since they can do damage even on a connection, but people didn’t play with them because it slows things down too much.
Plus, armor CAN simply stop damage. Armor was generally not built with flat surfaces (certain helms during the Crusades aside) because curved surfaces increase the odds that a blow will glance off you. A blade can bite into squishy flesh if you strike an unarmored opponent at an angle, but will slide off harder metal surfaces.
It’s also worth noting that in general when you’re fighting someone with armor that someone isn’t going to be standing around. They’re going to move. I’ve seen men in full plate do cartwheels. Never saw a woman do that, but I HAVE seen them pull off quite impressive Gumby-style dodges that the average person would consider impossible. And one of the things you learn is how to take a blow so it doesn’t hurt (because pain is a fantastic teacher). So someone relatively skilled with armor won’t be encumbered by it, and will have the knowledge and muscle-memory to use the armor actively, not merely passively.
As for soak, it never made much sense to me. I’ve seen people injured wearing armor (for that matter, I was injured wearing armor), and the armor was fine. In my case the injury was something the armor wasn’t intended to defend against–compressional force applied longitudinally to the clavicle. In other cases, the armor allowed force to transmit through it and into the person wearing it. The force of a war hammer or mace isn’t just going to stop, and the reduction in acceleration (or deceleration, if you prefer) from a gambeson can only do so much. Overwhelm that and you can still easily find yourself in a situation where the armor is more or less fine, but you are definitely not! Plus, if you want to get really technical, armor is never one piece; each piece would have a soak pool. This can work–Morrowind had a lot of fun with it–but most players aren’t prepared to learn the names for all those armor pieces, nor are they prepared to track that sort of damage carefully.
Bear in mind this all relies on the enemy being stupid. In fact, many fighting styles AVOIDED the armor. German longsword manuals taught people to find the gaps in armor (when they cared about kill the opponent at all, which wasn’t always the case–disrupting formations was easier and more effective). You can have all the soak you want in the armor, but if I’m sliding my blade up through the gap between your breastplate and gourget it’s not going to help you very much. Or I could aim for somewhere that doesn’t have armor to begin with. Fighters don’t just flail around wildly and hope to hit something; they spend a tremendous amount of time training to hit specific spots–and spent FAR more in the past.
If you like the storyteller system you should check out Burning Wheel! There’s a free PDF of the basic rules here: https://www.burningwheel.com/burning-wheel-gold-revised-hub-and-spokes-pdf/
The full game includes a rock-paper-scissors-y melee combat system that has weapons specialized for countering armor like you mentioned at the end there, and the armor type are closer to your list. Not saying you would want to throw out your whole system and switch but it has tons of ideas that I think would interest you.
For a good, illustrated discussion of the armors mention in the original D&D, see this post over at Zenopus Archives:
https://zenopusarchives.blogspot.com/2016/12/gygaxian-armor.html
It includes an illustration from the obscure Dungeon Master’s Adventure Log, which is pretty much the only illustration for what these prototypical armors were supposed to look like. (I say prototypical because something like “plate” covered a wide range of possible armors, which were all more-or-less equally effective, hence “armor class”.)
These are all supposed to be historical, but are based on only the best scholarship available to a 1970s Wisconsin public library. For things like split, which is described as “light chain, greaves, and a leather coat into which are laminated vertical pieces of plate with shoulder guards”, I suspect the text hews very closely to some source text, then the artist had to just make the most sense of it they could.
The source text is almost certainly Ffoulkes, as mentioned in that (lovely) link you provided. From that description splint is obviously a brigandine, let alone when you add in the illustration. Ffoulkes, who Gygax cited, specifically mentions how they’ll evolve into brigandines.
Ringed is on like the fourth page of Ffoulkes, and it is hard to judge Gygax for being misled by it. Banded is also discussed there, and is…not what Gygax produced. It also doesn’t make much sense because Ffoulkes himself calls banded mail bullshit (and really heavy).
It is notable that Ffoulkes clearly states how theoretical all of this is, and Gygax appears to have basically gone “we’ll throw it all in” rather than go try and find a book more recent than sixty years old.
Ffoulkes also talks at length about how leather would be studded without ever clarifying in detail what this means, and so Gygax presumably did his best with what he could. There’s an illustration at figure 9 of Ffoulkes with an example of studding on the legs of all places, which suggests it’s either decorative or (at any rate) not just a misinterpretation of a brigandine.
Also I am genuinely shocked a man who wanted different rules for every polearm didn’t create different rules for every helmet. Ffoulkes has a beautiful selection.
“Splint armor, which is, as the rule-book says, “narrow vertical strips of metal riveted to a backing” of some flexible material, was never a complete armor, but a form of protection for the arms and the legs…”
Splinted torso protection was worn by the Pacific Northwest Indians, but it was made out of wood, maybe bone, not metal. Again, it’s in Stone’s Glossary.
AFAIK, the point of vest of vertical splints was to keep arrows out of vital organs. When shooting a human with a stone-tipped arrow, the arrowhead has to be aligned in a horizontal plane, to have a chance of passing between human ribs. Putting a barricade of vertical sticks in the way makes this much harder to do. Yes, they could use hunting arrows, where the arrowhead is in a vertical plane to pass between deer ribs. This could pass between the splints, but then, it would hit the target’s ribs and stay out of lungs or heart.
Unlikely, because the arrows fired by most peoples, most places, spin in flight to stabilise them so there’s no way to guarantee what angle the arrowhead will hit at. (Possibly a high velocity modern compound bow at very short range won’t have time to spin.) Not an expert on native American weaponry, but I doubt they would be any different.
Vertical splints because branches are mostly straight, so fit more neatly to the human torso if arranged vertically than horizontally?
Arrows might have spun incidentally, but they weren’t spinning nearly fast enough for stabilization. Stabilization was done by the fletching.
Couldn’t people have simply used piercing arrowheads that would fit through the gaps regardless of orientation to get around that? All that really requires is sharpening the end of a stick (though I’m sure a bone/stone/metal arrowhead could be fashioned as well). Sure, a cutting arrowhead will do more damage overall if it hits, but if it’s a choice between “bouncing off armor” and “an arrow sticking out of a guy’s torso” it seems like an obvious choice…
“AFAIK, the point of vest of vertical splints was to keep arrows out of vital organs. When shooting a human with a stone-tipped arrow, the arrowhead has to be aligned in a horizontal plane, to have a chance of passing between human ribs. Putting a barricade of vertical sticks in the way makes this much harder to do.”
This sounds very unlikely for the reasons other commenters have already given, so I’ll just add this: if your armour material of choice is the stick, that is a long, thin, straight thing that is tricky to bend. If the thing you’re trying to armour is a human torso (or indeed a human limb), that is a roughly cylindrical thing with a vertical axis. If you want to wrap a cylindrical thing with a vertical axis using a lot of long thin straight things that are difficult to bend, well, there’s really only one way to do it…
A lot of confusion stems from the fact that, especially in original D&D, with combat rounds of 10 seconds or more, an “attack” isn’t just a single swing. It means a whole series of attacks, parries, ripostes, etc. Armor class doesn’t measure how well a given kind of armor deflects a blow, it’s more or less an actuarial table comparing the “survivability” of warriors in different kinds of armor.
I remember getting into a discussion with a friend about rounds of combat. Another person had told my friend that 6 attacks per round was unrealistic (too fast). She asked me what I thought and I responded “Yeah, any reasonable fighter should get off way more shots than that!” Setting up to throw that first shot involves a lot of standing and apparently random twitching and feints and whatnot, but when it comes to actually striking blows sword-and-shield fighters are taught to use one blow to set up for the next, in a flurry that overwhelms the opponent’s capacity to counter. If you’re only taking one shot a second, you’re either being SUPER precise (not a bad tactic if you can get away with it) or are SUPER slow (which is absolutely a bad tactic). Add the fact that many fighting styles were single-time, where the defense was an attack, and it’s even worse.
Further, the reality is that most sword fights are going to last a few seconds. Again, the setup can last a lot longer, but once you start trying to hit each other, one of you will win and one of you will lose after the first minute or so–or you’ll disengage and set up for another round or try to get out of Dodge. Anything that lasts longer than a minute or two isn’t a fight, it’s a complicated dance number (which were actual things in the Middle Ages).
So for my part, I view “rounds” as more or less a necessary though abstract thing. You’ve got to control the chaos of combat somehow, and have everyone able to understand what’s going on. Initiative and rounds are as good a way as anything else. None of it’s realistic anyway.
Fights can last longer than a few minutes…if you’re wearing enough armour.
Of course, that’s when you grapple.
Excited to hear you’ll be tackling leather armour down the line! Just wondering, have you stumbled upon these at some point?
* Jackmeister’s YouTube video on Mongol heavy cavalry armour, including the use of leather by steppe-adjacent cultures (I’d also recommend checking out his series on nomadic blacksmithing, if you have the time): https://youtu.be/WuDE6y6mJ4k
* This paper from last year on a 8th-6th c. BCE leather scale set from Xinjiang that apparently matched the style of contemporary Neo-Assyrian armour (which probably says some interesting things about the scale of cultural exchange through the Inner Eurasian steppes): https://www.sciencedirect.com/science/article/pii/S1040618221005553
If you were a geologist would you critique the Underdark this harshly? I detest the use of medieval historicity in this way as a limitation on fantasy, instead of what it is – a convenient shorthand for an aesthetic. This reeks of Tolkien vs. Lewis regarding lampposts.
“If you were a geologist would you critique the Underdark this harshly?”
I’ve seen a geological analysis of Skyrim. So yes. Geologists absolutely would this. (Hint: Bioturbation covers many sins, especially when you have something the scale of a purple worm!)
I’ve also done ecological analyses on alleged cryptids, mostly because I was bored and it was fun. For that matter, I’ve analyzed the stratigraphy of a Snickers bar (to teach several of the principles of stratigraphy–people remember geology they can eat!).
The thing is, none of this is meant to detract from enjoyment of the game. It’s meant to be part of how we enjoy the game. The conclusion isn’t “…and therefore this is a horrible game and you should never play it”, but rather “…and this was a fun exercise that I enjoyed a great deal and I hope you did too.” It’s a shared starting point, a point at which someone that’s an expert in a field can use to launch into a deeper understanding of that field. And there is such a thing as geeking out. Some people are passionately devoted to their chosen professions, and take sincere and deep joy in exploring the implications thereof with other people. The intent isn’t to bash on the thing, but to share one’s joy, even if it means pointing out flaws in something.
tl;dr: For some of us, this is how we have fun. You’re literally doing what you’re accusing Bret of doing.
…Veins of the Earth is basically “what if the Underdark, but informed by actual caves and caving and the extreme limitations of an environment that’s quite that hostile.” It is extremely popular and widely viewed as a masterpiece despite the bit that it needed like a zillion house rules to be playable and everyone knows that.
Perfectly reasonable basis to critique things on. Accuracy generally creates verisimilitude (though not always, it is a balance). Verisimilitude makes things more gameable.
The best definition of HP I ever saw was Kevin Crawford’s in his most recent game (Cities Without Number).
“A person’s nearness to defeat is measured in Hit Points.”
He also has melee weapons do “Shock” which is automatic damage done to targets with an AC below a certain number.
I do feel like there must be a better treatment of armour class, but the question is how. I feel like the Storyteller system method
Also, “heavily modified Storyeller” is code for “Exalted” which is code for “weeb” which all raises the question of who Bret Devereaux thinks is best girl?
I liked the Wounds/Vitality system from SW d20 and 3e UA: Vitality is calculated exactly like your HP would be. It represents fatigue, luck, near misses, and scratches/glancing blows… not “real damage.”
A critical hit is “real damage”, and it doesn’t do extra vitality damage. It goes straight to your Wounds. Wounds (before you take damage) equal your Constitution score… and only go up if your Constitution goes up.
If you run out of Vitality, you start to take Wounds damage. If you run out of Wounds damage, you’re dead. So a crit against a moderately healthy opponent can kill then no matter how high level.
But mainly: that definition of Vitality. That’s what is in my mind as the definition of hit points. It’s not real damage, it’s near misses, luck, glancing blows, fatigue due to lots of dodging, etc.
I think HP really started to struggle when people started using it for things like falling damage and poison without taking into account how to resolve that.