Hitpoints don't represent anything, actually

Daniel Norton of Bandit's Keep beat me to my last post with this video, and I didn't even realise! And sure enough he's beaten me to the punch again - although in fact I managed to see the video first this time. This post can partly be read as a response ot his video. It's what you might call a long read, so grab yourself a cup of tea.

The topic of hitpoints in D&D is one of those things that seems perpetually to be puzzling. Everyone has a different take on what it is that hitpoints represent: Are they bodily toughness, exhaustion and fatigue, combat awareness, luck, some combination of the above ("hit protection"), or something different entirely?

But the setup for this whole debate hides an assumption, and one that we need not make. Let me propose something radical: Hitpoints don't represent anything, actually. That is, there is no quality that your character posesses - be it something as abstract as luck or as concrete as what you might call corporeal structural integrity - that your character's hitpoints represent, that they track, that goes up and down in rough proportion with your Hp.

Allow me to explain (said the Lawful Cleric caught with the succubus).

🕷

I've played at tables where hitpoints are literally treated as the toughness of your body. These are the tables that tend to have characters sustain what are described as grievous injuries that nevertheless seem to get shrugged off and forgotten about moments later. They also tend to be the tables with NPCs suffering from "Kevlar Neck Syndrome", a condition where, should you manage to get the drop on an NPC and slit their throat from behind, they usually turn out to have an esophagus made of whatever it is they make those police dog training suits from.

We're familiar with the problems with the corporeal structural integrity approach to hitpoints. Briefly, in diegetic terms, nobody has a body so tough that they can survive a sword thrust with no ill effects, let alone falling a hundred feet (a mere 10d6 Hp damage) or being immolated by a dragon (varies). The fifth edition D&D Dungeon Master's Guide, which seems a particularly bad offender for this, offers guidance that wading through a lava stream should do about 10d10 Hp of damage (average 55) per six second round. That means your 12th level Fighter-Champion, whose main distinction is that they're really good at hitting things with a sword, can spend two or three rounds dithering about which side of the lava stream they want to be on, without any immediate ill effects (and then regain their lost Hp swiftly after, because of the 5e resting rules).

[The maths: Hp from levels = 76 (full hp at first level, assuming taking static number afterwards instead of rolling), CON bonus assuming (conservatively) 16 CON at 12th level = 36, Second Wind ability grants avg. 17.5 for total avg. Hp before incapacitated = 129.5]

"I can do this all day."
Art by Jeff Easley

The problem here is not even so much that this isn't realistic. I don't want to get into the whole realism/genre discussion here - the real point is that this sort of incongruity affects players' abilities to make decisions, because they don't know, from one moment to the next, how tough their characters actually are. Should we be scared of that hundred-foot drop? The book says my character can survive that. Why is this two-ton stone block trap deadlier than the six-ton-flying-tank-with-a-flamethrower of a dragon that we fought in the last encounter? An assassin is tracking us, armed only with a knife - should I be scared or not? Bonus round if the GM uses the 5e designers' own workaround for the Assassin subclass, namely having the character simply do loads more damage with that knife: Is this one of those hyper-deadly NPCs or not? How can we tell?

(Having said I want to avoid the genre discussion, I will just briefly note that nowhere do Conan, Aragorn, Fafhrd etc. ever roll up their trousers and simply wade through lava, for much the same reasons of consistency I've mentioned; once a character has shown that they can do that, the reader will look incredulously at any purported threats or perils the author subsequently produces.)

In old school D&D, the hitpoint problem is in some ways less acute, and in some ways more. At lower levels, hitpoints look to be more in line with this sort of "body points" conception - your first level fighter has d8 hitpoints, and a successful hit with a sword will do d8 damage, for a rougly 50% chance of taking them out. Recovery times for hitpoints, orders of magnitude longer than in the modern game, seem to speak to the idea of recovering from actual physical injuries rather than depleted stamina. But the same problems persist. In particular, the old school style of play is much more interested in things like traps designed to do obvious physical damage. But, importantly, hitpoints in D&D don't provide for this. As many have noticed, a character is normally fighting at peak effectiveness until they drop to 0 and die. But this doesn't sit well with the game as played. You really did fall in that 10' deep spike pit - the fact that you have 5hp left afterwards doesn't signify that it was a "near miss" that merely depleted your "luck". So where's the bleeding and the broken bones?

🕷

Most popular attempted solutions accept that hitpoints are in some way abstract - that can hardly be denied, given the problems with the "body points" view. As I say, there are a few answers to what, therefore, hitpoints actually represent. What I'll try to do in the next section is to show, briefly, why they can't actually quite be any of these things. First, though, I'll briefly give my proposal, as it will be helpful to refer back to it.

Hitpoints don't represent anything, actually. There is no quality that your character possesses that hitpoints track, that varies even roughly in line with them, and is exhausted when they're gone.

Hitpoints instead embody a simple convention. The convention is this: If your character loses all of their hitpoints, they die. (The latter modulo special rules like death saving throws etc.) When your character experiences something that could kill them, especially if that is in combat, the GM will usually deduct from their hitpoints. Note this does not have to be something that inflicts harm or puts them "closer to death" than before - hitpoints aren't tracking your caracter's "health", in any sense. The GM can bypass hitpoints to inflict harm on your character, for example, by saying that they are fatigued, are poisoned, or lose a limb. This does not have to have an effect on hitpoints, although it can, at the GM's discretion.

I'll just say it again for those in the back row: Hitpoints represent nothing, are nothing more than a simple "if/then" convention: If your character loses all of their hitpoints, then they die. 

🕷

One very popular approach to the hitpoint problem is to see hitpoints as a sort of gauge of stamina or exhaustion that, when depleted, leaves your character open to a killing blow. This has an old pedigree: Ken StAndre and Liz Danforth suggest exactly this view at least as early as the 5th edition of Tunnels and Trolls, which hails from the 80s (in T&T, damage comes directly off your CON score, which effectively serves as your hitpoint total - atributes can be increased through play, sometimes to many times their original score). On this "stamina" view, your character generally suffers nothing more than cuts and bruises before the final blow that drops them to 0hp, which is a substantial one.

But (and you'll be hearing this again) this can't quite be what hitpoints are in D&D. We can see this in how hitpoints are used: While hitpoints don't properly represent corporeal structural integrity - this is the thing the "stamina" answer gets right - neither are they used as a measure of endurance or stamina. If that were the case, we'd expect to see some sort of test against hitpoints when characters undertake feats of endurance, such as trekking long distances or climbing a mountain. But D&D already has a statistic that's used to test stamina: Constitution. This is where comparison to Tunnels and Trolls is actually helpful: If D&D's hitpoints were a sort of stamina bar that degraded in combat, D&D's hitpoint system ought to work like that in Tunnels and Trolls, where a character's Constitution attribute rises and falls as their stamina depletes and recovers. But in D&D, hitpoints are a separate gauge, one that only deplete when characters "take damage", and only recover when they "heal". D&D does in fact have systems for tracking degrading stamina. In old school D&D, characters that don't sufficient spend time resting while in the dungeon or travelling in the wilderness suffer penalties, such as a cumulative -1 penalty to all rolls, depending on edition, while modern D&D has a bespoke "exhaustion" status effect for when characters' stamina is depleted. Neither of these approaches interfaces with hitpoints.

I'm aware that people may not yet be convinced, so let me offer another argument: Just as your character's body doesn't become physically tougher as they level up, at least not to the extent suggested by the "corporeal structural integrity" model, neither does your character's stamina double or triple as they level up. Your character, who can manage a six day hike without rest at lv1, doesn't become able to manage a twelve day hike at lv2, and so on. Yes, their saving throws (which the GM might have you roll against to counter exhaustion) improve, and yes, if you're playing a newer edition, their CON score might improve, and these things represent their stamina improving. But these things are not their hitpoints.

One issue here is the role of Constitution in affecting hitpoints. The fact that CON factors in at all implies that hitpoints have something to do with stamina or hardiness. But, as we've seen, taking that relationship to be particularly clear-cut poses problems. But even though they are related, note that hitpoints are downstream of Constitution; you have (slightly) higher or lower hitpoints because you have more or less stamina, not because they're the same thing.

🕷

What seems to be the relevant difference here is that, while stamina is the concept of how tired your character is and how well they resist fatigue, hitpoints only usually get invoked when there's a risk of death, usualy in combat. So maybe we can generate an answer out of this observation.

Lo and behold, people have. Games like Into the Odd and its many descendants offer the idea that hitpoints should really be thought of as "hit protection". For example, ItO separates out your hitpoints from your "meat points", in this case your Strength score - you only start taking damage to the latter when you run out of the former. This is reflected in different recovery times for each; in ItO and derivatives, characters recover Hp uncharacteristically fast, for an OSR game, typically regaining their full complement with a few minutes' rest.

I actually really like the hitpoint/meat points distinction. I think it successfully articulates a distinction that's very helpful for a GM to bear in mind, and can be helpful to have baked into the mechanics. You can also use the rule to differentiate between foes that are tough but not good at fighting/avoiding physical damage, and vice versa. I use the Into the Odd mass combat rules in my Swords & Wizardry game, which are exactly the same system but with divisions of troops instead of individuals, and at higher levels of abstraction, it works wonders.

So what does "hit protection" represent? Well, usually it's an abstraction, representing your ability to protect yourself from harm - a mixture of luck, resilience, combat skill (albeit wholly defensive) and my personal favourite, "combat awareness". 

Put a pin in that "it's an abstraction" - we'll come back to that again in a moment.

I usually introuce the concept of hit protection to players as "combat awareness", because this captures the general idea. It's partly a measure of your control of the battlefield - your ability to predict where that next arrow is coming from, and get your shield in the way. It's also a good measure of your character's experience; more experienced combatants will be better at seeing attacks coming, and retaining composure. The system works particularly well for modelling OSR-style combat, because hit protection effectively works as a gauge for how close the combat is to getting beyond the party's control. Once it ticks down, everything suddenly starts going drastically wrong quite quickly. It creates tension.

This can't quite be what hitpoints in D&D are representing.

The moves in the argument are effectively the same as before. Other systems than D&D have statistics for luck, which can be used to e.g.: mitigate damage, until it ticks down and can't offer protection anymore (my mind goes instantly to Fighting Fantasy and its full-fledged rpg counterpart, Advanced Fighting Fantasy). But again, hitpoints are not used like "luck tests" except perhaps in a very specific sense. That sense - of testing your luck to see if you survive the pit trap - we talked about above.

But even here, the way hitpoints are functionally used in the game is at odds with this idea. For example, if a character wants to "try their luck" by swinging from a chandelier, the GM might well ask for some sort of saving throw or ability roll to check if they lose their grip and fall. What they almost certainly won't do is automatically subtract from the character's hitpoints to represent them "depleting" their luck, and test whether they push their luck too far and fall, becoming combat ineffective. So it seems wrong to look at subtracting from hitpoints as modelling a straightforward "luck test" to see if you survive.

Combat awareness seems perhaps slightly closer. But again, this can't quite be what D&D's hitpoints are suppose to be. When the Cleric heals you, are they restoring your awareness of the situation, your control of the battlefield? Shouldn't a party or group of monsters that gets surprised suffer an Hp penalty, since their combat awareness is compromised? (Incidentally, this is something the Into the Odd system does model: The party that suffers an ambush can take hits directly to STR, or have their Hp halved for the first round, etc.)

A resource that takes weeks or months to regain through "natural healing" surely isn't measuring something like "control of the situation". What would that even mean? Again, this is why characters in ItO regain all their Hp after a few minutes' rest, on my reaing at least: The only reason you wouldn't stop to rest and reset your "control of the situation meter" is if the sitution is still ongoing, i.e.: you're being chased or attacked, or you're having to hurry along before reinforcements arrive without chance to regain your composure. There's also the fact that D&D characters don't have another layer of "meat points" underneath their hitpoints. Once your hitpoints are at zero, you're combat ineffective, even if not necessarily outright dead - even with death saves or negative Hp rules, there's no "up but vulnerable" or "injured but still effective" state in D&D.

So hitpoints aren't actually "hit protection". But the hit protection answer isn't a dead loss. Like the previous answer, it tells us something informative, namely that hitpoints are an abstract measure of (something like) closeness to death, specifically for use in situations like combat.

🕷

And here we come to the final argument. Because I'm sure a good number of people who have made it this far will be balling their fists and shouting something like "They're an abstraction, you dummy! They just represent how close you are to dying!"

I hear you. At a certain point, I agreed with that view. D&D contains loads of abstract quantities, such as levels and experience points (and in latter editions advantage, sorcery points, inspiration, spell slots and so on and so forth). It wouldn't be surprising for hitpoints to be on that list.

But lets take a step back for a minute and discuss what it means to be "abstract". I promise it'll be brief. Because "abstract" gets used imprecisely, both in gaming and elsewhere. Questionable definitions for "abstract" I've seen range from "something you can't see or touch" - often true, but not stictly accurate - to (shudder) "something fictional that doesn't really exist as a part of reality".

This is not strictly what the word means. An abstract property is one that isn't instantiated by any single quality in the world, but can be instantiated by many. For instance, the property "redness" is abstract because there are many ways of being red - many different substances a thing can be made from that make it red, many different ways of producing the correct frequency of light. Likewise, you can have abstract objects. The number three is an abstract object, not because you can't see or touch it, but because there are lots of ways of having three things - you can have three cups, three dice, three arguments, and so on. This is actually where the word comes from: Humans see these different instantiations, and we abstract (verb) the property out as a thing separate from its instances. Abstract properties are the product of a process of abstraction.

So what does it mean to say that a quantity in D&D represents an abstract quantity in the game world? Well, just that: The quantity is capable of being instantiated by multiple things. So, for example, hit protection in ItO and its hacks might represent a combination of how lucky you are, how fast your reflexes are, how experienced a combatant your are, and so on. Abstract properties are real - red things really are red, your expereinced ItO character really is harder to hit - they're just, well, abstract.

The point is that D&D hitpoints don't really represent any property, even an abstract one. Or, if you like, the thing they "represent" is so abstract as not to be a property in the game world anymore, but simply a game structure. Remember all the times above when I pointed out that hitpoints could almost be something specific, like stamina or luck, except that we don't really use them like that anywhere outside specific situations, like combat and taking damage from a trap? This is because what hitpoints really have is a game function (stopping your character from dying), keye to activate in those specific situations. When we think about what they represent, we start casting around for anything in the game world that might be causing it to behave in a way that matches the game mechanics. But whatever we might think of for the specific situation - stamina, luck, reflexes, awareness - hitpoints don't track that thing, because we never use them to represent that thing outside of combat, or traps, or other situations where you "take damage" - whatever that's suppose to mean here.

To put the point another way, consider that we could take all the properties that hitpoints could ever represent in any conceivable situation and bundle them together under a label like "death-defyingness". This is just a way of noting that hitpoints track nothing more than how close, in a mathematical sense, your character is to death, whatever that means in the specific situation. But this is just to point again to that same game structure - to say "Hitpoints just represent how close your character is to death" is just to repeat the convention from earlier: If your character loses all of their hitpoints, then they die.

Abstract properties like redness or luck or combat experience exist within the game world that you interact with when you play an rpg. If hitpoints did represent anything, it would be one of these abstract properties - something in the game world. But hitpoints in D&D really just come down to this game structure, and structures of the game do not exist in the game world. They aren't like a character attribute like Strength or Charisma that tracks a property of your character. Hitpoints only "activate" in certain situations - situations where it helps the player to be able to get a handle on whether the GM is going to rule them "dead" or not. That function is al there is to them.

In game design terms, it's hard not to see the bundle of confusion that is hitpoints being inherited from wargames. In Chainmail, for instance, a Hero takes four simultaneous hits to kill. This is not because they're physically four times as tough, nor does it "represent" or "track" the enemies wearing them down, in any meaningful sense. It's just that it would feel wrong for Conan or Aragorn to be killed as easily as a regular infantryman, so they get a free pass for the first few hits - it's as simple as that. Wargames are generally both more abstract (there's that word again) and have narrower scope than roleplaying games, not generally dealing with situations other than battle of some sort, so the question of what these "extra hits" represent, if they even represent anything at all, doesn't generally come up. It's a different outlook. Fundamentally, they're not there to represent anything; they're a game structure that's there to get a game result that feels right. Because ttrpgs typically deal with characters and their characteristics in much more depth, we're primed to look for one-to-one representational relations. But the provenance of hitpoints as a game structure simply gives us no reason to expect to find that in this case.

🕷

This has gone on long enough. Can we salvage a gaming lesson from this, admittedly wholly negative, argument?

Well, knowing something new about how ttrpgs work (and I do think the arguments here furnish us with some knowledge) always puts us on a footing to run better games. For one thing, we might know now not to describe a characters' loss of hitpoints in misleading ways. I tell my players not to worry about what hitpoints represent. This goes some way to averting those "Wait... what!?" moments. For example, it helps if players know that things that would logically instantly kill them - 100' drops, lava, two-ton falling stone blocks - will kill them regardless of how many Hp they have left.

More broadly, this gives the GM some more freedom to treat the characters as capable of suffering physical harm, something that, as noted, doesn't generally actually happen all that often if you insist on modelling it via hitpoints. If hitpoints are decoupled from physical harm, you can deal a grisly injury to a character (perhaps when an enemy rolls a natural 20) without having to render them combat ineffective. This gives that state of "injured but still fighting" that the hit protection/meat points system has, without having to worry about rewriting the rules. I actually prefer the D&D version for the individual level and the Into the Odd rules for the division level, because at the individual level I feel like the line between "I'm fine" and "I'm injured and everything's spiralling out of control" is thinner for an individual than for a division. While you can watch a phalanx of spearmen gradually accrue fatigue and low levels of casualties, for the individual everything's going generally okay until it's suddenly really not.

This dovetails nicely with my own death and injury rules. Never mind what hitpoints represent - what you know about the situation when someone's lost their last hitpoint is that they've dropped out of the combat. So here is where I start getting the players to roll on tables to see if they've taken an arm wound and are hanging back, or have suffered a blunt instrument brain-ectomy and are currently spectating from another plane. As I note in the linked article, this adds a nice rhythm of tension - indeed, one that feels similar to the controlled/uncontrolled rhythm of the hit protection/meat points system.

I suppose the positive upshot of accepting that hitpoints don't represent anything, actually, is that you no longer have to worry about what it is hitpoints represent - you no longer have to justify particular rulings to your players, or worry about the inconsistencies that get generated, or whatever. I find that players in this mode ask questions like "Will this thing instantly kill me?" To me, that seems to signify that they're not taking the game structure for granted, and are maybe defaulting to in-world thinking, which is of course always desirable. Certainly I don't usually get that response of "But how did that assassin kill me? I have twenty hitpoints!" And of course my NPCs don't have Kevlar Neck Syndrome, so that's a positive. I find that players who come from the play culture of modern D&D have reactions of pleasant astonishment when they manage to kill a foe outright with only a logical argument - that, to me, says the approach is doing something right.

I don't expect many will agree fully with this. I look forward to finding out.