Rob: We had a few design elements in 4th Edition that took shape early and maintained that shape. Death and dying wasn't one of those elements. We tried at least four different death systems with vastly different consequences.
Enough disagreement arose about the goals of a death-and-dying system that we held meetings to negotiate what we wanted. Some people argued that death was passé -- the example of online RPGs in which characters respawned indicated players wanted to be free of death's threat. Most of us didn't buy that reasoning. We knew that online RPGs had their own separate logic. Our game required a realistic possibility of defeat and death to maintain its excitement and dramatic tension.
Heroes are defined by the villains they face and the challenges they overcome. Take away a plausible threat of death and failure, and heroes become boring automatons. So what we needed were game mechanics for death that worried players, keeping them on their toes. But the mechanics still had to be fun at the table, and they couldn't take players out of the game too often or for too long.
No one looks forward to his or her character dying. But when it happens, it had better be memorable and offer glimmers of hope. It has to involve something other than crushing despair.
James: Early on, Rob, Andy, and I sketched out some ideas about how death's impact on the game might vary over the course of the three tiers of play. At the lowest levels of the game, we figured, death could very well mean making a new character. When you've only played a character for a couple of levels, you're not that attached to it yet. Making a new one isn't that terrible a price to pay for an encounter gone terribly wrong. By the high heroic levels, and certainly the paragon tier, death comes with a more significant cost. It means stopping your adventure for the day, paying the price, and possibly story-based implications, up to and including failing a mission. At epic levels, though, death might become a speed bump. The consequences of failure in an encounter or an adventure might be much more significant than the life or death of a single character. When the fate of the world or even multiple planes hangs in the balance, fates far worse than death can await characters who fail.
RH: Despite our intentions, the earliest 4th Edition death mechanics erred on the side of dispelling all of death's threat. We actually ran playtests in which no one died. That was mainly because we'd wimped out in our design, and the only way a dying character could really die was if the DM actively targeted the character to finish him or her off. Of course, that wasn't a style of play that anyone could become comfortable with. We couldn't make character death a game element the DM had to execute intentionally that way. We'd be forcing an unpalatable choice between hurt feelings and a lack of dramatic tension.
Character death in D&D has to come as a natural and even expected consequence of taking massive amounts of damage. The system needs to roll you toward that endpoint. Death mechanics in 3E had a decent handle on that. We couldn't be satisfied with a system that failed to improve on 3E.
JW: All of our various stabs at mechanics were aimed at a single goal. As Rob pointed out, some aspects of 3E's death mechanics handled the goal well. What we wanted was a specific dynamic at the table -- an atmosphere when a character takes enough damage to go below 0 hit points.
The entire group of players takes a breath and holds it. Things get tense. Everyone shifts into emergency mode. The group recognizes that something needs to be done, and quickly. Characters do their best and most heroic things at these moments.
There's a second part, though. When the character gets back up, there should be a sigh of relief. The worst is over, but the character isn't out of the woods yet. On one hand, we don't want the punching clown effect, where one hit takes the newly revived character back down. On the other hand, we don't want the character back at full strength, ready to take on the world again.
Save or Die
RH: Third Edition's save-or-die effects were one of the major influences on our philosophy of character death in 4th Edition. We referred to any effect or attack that could take a PC out of the game for more than a round with a single die roll as a save-or-die effect, even when the consequence wasn't death. Paralysis, confusion, stun, and charm -- all these 3E effects frequently functioned as save-or-die effects. They made a player's enjoyment of an encounter, or an entire evening, hinge on a single saving throw.
I thought of save-or-die effects as one of the subtle game design problems that were capable of killing entire 3E campaigns. When one or two 3E PCs couldn't fail a particular saving throw without rolling a 1, and the other PCs in the group could hardly make that same save, campaigns had a way of stuttering and halting. From the beginning of our work on 4th Edition, we knew that we weren't going to use save-or-die effects that could take out PCs with a single die roll.
One of our earliest alternatives to save-or-die effects was a sort of tracking system. This system measured a character's progress toward various debilitating states by sliding characters up and down on numerous, specific tracks. Each track's endpoint could result in a condition such as blinded, petrified, or dominated.
The track system ended up working very badly for conditions. That didn't prevent us from attempting to use it for death and dying. For a while, the phrase "I'm on the death track" meant that your PC was sliding toward doom.
A death-track system never had fans, but it knocked us out of thinking only in terms of negative hit points that dropped by one each round. That was a habit surprisingly hard to break. Much later, the track mechanics found a decent home in the disease rules.
Fighting the Negatives
RH: One of the reasons we tried so hard to make the track mechanics work was some designers thought that using negative hit points was unnecessarily complex. Players didn't need to be bothered with negative math. During several phases of design, we were consequently using systems that stopped tracking hit points when characters dropped to 0.
I was never convinced by the arguments against negative hit points. As we tried multiple systems that avoided negative numbers, I became entirely persuaded that a system based on hit points should keep using hit points to track the possibility of death. The final rules for dying at negative hit points equal to your bloodied value are exactly what I wanted.
Characters seldom die this way. However, if you're caught in too many breath weapons, or that scary brute crits you, you just might die from losing too many hit points. That can happen much faster than you can fail three death saving throws.
Save Against Death
JW: Speaking from personal experience, when you're having an off night with saving throws, it doesn't matter whether you fail your death saving throw or the one against the ongoing acid damage from the black dragon's bite attack. Six of one death, half-a-dozen of the other … .
RH: The move away from the condition tracks led to organizing the game around effects that lasted a round or that lasted until the PC made a save. We saw that the saving throw worked and was a fun way for a player to rid a character of a condition. That made me sure we could use the saving throw to keep track of how long PCs hang on while dying. And better yet, we also used it to give PCs a chance to get back into the fight even if no one managed to heal them.
There's always hope for a dying PC. Even when your character is dying, the death saving throw you make at the end of your turn has at least a 1 in 20 chance of getting you back on your feet. As a result, a player stays involved with the game even when his or her PC is down.
This has everything to do with the psychology of victory over incredible odds. Victory after you've been dying and failed a death saving throw or two is even sweeter than an easy win. It's the type of miraculous comeback that fantasy fans expect from their heroes.
JW: The death saving throw system works well within the framework of the goals we set for ourselves. When a character drops, a very scary timer starts ticking that really does create the sense of emergency I was talking about before.
When you get back up, you typically have enough hit points to survive another hit, because healing starts from 0. Those negative hit points don't count against you. But your failed death saving throws are a black mark on your character sheet until you take a rest. If you do go down again in the same fight, you could be in trouble.
One less obvious factor is that healing is a limited resource in 4th Edition. Cure light wounds spells in potions or wands meant healing never could be as limited in 3E. When I finish a fight and realize it took three-quarters of my healing surges, it feels like a much tougher fight than the one that ate up ten wand charges in my last 3E game. And when I go into a fight with only one or two surges left, I play a lot more cautiously. I know that I could drop and not be able to get up.
RH: James mentioned before that epic adventurers occasionally face consequences worse than death. When epic characters fail, it's not necessarily those characters themselves who die. Instead, the consequence for epic failure can be that big parts of the world the player characters were trying to protect end up dead. Bringing them back to life isn't as simple as using a ritual called Raise Dead.
JW: Whatever way it happens, death is inevitable. Sooner or later, player characters die. Then the game needs to have an elegant and believable way to handle it.
Coming back from the dead was punishing in past editions -- from losing a point of Constitution to losing a level. The logical consequence of that punishment is that players preferred making up a new character to raising the dead one. In that environment, character death means you can kiss your campaign's story continuity good-bye. Or you can welcome a new character named Bob II who's virtually identical to poor, deceased Bob.
The Raise Dead ritual is the normal means by which characters return from death in 4th Edition. We couldn't find a good reason to keep resurrection and true resurrection in the game, so we just raised the price of Raise Dead for higher-level characters to ensure that it remains relevant as characters advance levels. Raise Dead is also a ritual that takes eight hours, so it definitely means that the party is taking an extended rest. That could have its own consequences in the adventure.
The ritual also comes with a "death penalty" that's like a negative level in 3E. You take a -1 penalty on just about every d20 roll you make, but that penalty goes away after three milestones. It lasts long enough to make sure you wish you hadn't died but not so long as to severely hamper you or make you wish you'd just made a new character.
For epic-level characters, though, it can still be true that death is just a speed bump, at least once each day. Most epic destinies give characters some way to cheat death. This led to that choice phrase in those epic destinies, "Once per day, when you die," in which I personally take great delight.
The paladin's gift of life power (level 22 utility) can also bring a character back in the middle of a fight. I'm sure that's not the only epic-level power we're going to publish to allow instant resurrection.
About the Authors
Rob Heinsoo led the design of the 4th Edition D&D Roleplaying Game and currently serves as the Lead Designer for Wizards of the Coast Roleplaying R&D. His 4th Edition design credits include Martial Power and the Forgotten Realms Player’s Guide. His other game designs include Three-Dragon Ante and Inn-Fighting.
James Wyatt is the D&D Design Manager for Wizards of the Coast Roleplaying R&D. He was one of the lead designers for 4th Edition D&D and the primary author of the 4th Edition Dungeon Master’s Guide. He was one of the designers of the Eberron Campaign Setting and is the author of several Eberron novels.