How to Implement Save or Dies in a Balanced D&D System

[Foxwarrior]

If the goal is to present death as a threat without having it actually happen much at all, at least exploit players' loose understanding of probability by making death involve a check that can only be failed on a natural 1 (on a d100 if you prefer). That way you can point at it and legitimately say there's a chance without the chance being so high it actually happens to someone in every campaign.

That said, my preference is to not play out fights that are so meaningless it wouldn't be interesting for the players to have losses to them, and to offer a variety of revival-but-with-some-change-you-can't-completely-ignore options for players that want to keep playing a lost character, like emergency prosthetics or the character being replaced by a creepy shadow clone (see: Gandalf)

[Neo Phonelobster Prime]

...at least exploit players' loose understanding of probability...

Two things. No. Wait. Three things.

1) Not relevant to death flags. But hey, maybe it's time to back away from that topic and shut it's proponents into a small padded room.

2) If your players understanding of probability is so loose you can make them roll a d100 tell them their character dies on a 1 and then they celebrate victory against the odds... that's a REALLY loose understanding of probability.

3) That's just normal game rules. Pretty much all of them. That's just what they do. Breaking fights into turns and actions and attack rolls and damage rolls and every other mechanic. ALL of that already exists. And it's a hell of a lot better in it's more standard many layered complexity at obfuscating from players that actually in net the odds actually probably ARE in their favor than just straight up rolling a fucking 1 on a d100.

We can argue endlessly about how many layers of mechanics could or should exist between success and defeat states, we can argue about the odds. Because there isn't a single right answer because genres and preferences.

But the thing is we weren't just now having an argument about the layers of mechanics between these states. What we were actually discussing with death flags and the retroactive soft balling of defeats alternative was about arbitrarily redefining the defeat states themselves.

An additional "stabilization" roll is not the same ball park.

I don't really care what number you want to roll on what dice as an extra roll you've just now decided to insert between your player and their character's defeat state. Especially if I have no idea what your other game mechanics are. I care about how you handle that defeat state and if it matches your actual design goals.

Trying to take that layer away (or not) in return for a +3 to attack (or not) was the whole discussion for a while now.

Instead just jumping in between that layer for an additional 99 out of 100 dice roll... frankly is still stupid. Its kinda different stupid. But IF you feel you cannot rely on the rest of your game mechanics to generate defeat states of a reliable safe severity for your preferences is it somehow so fucking hard for you to just say to your player "how much do you feel like that just killed your character?" while also being so apparently easy to deceive them into thinking that 17 they just rolled on that d100 was a "near brush with death" that somehow really gave them that engagement that was missing from your game?

Does rolling a 99% chance of success death stabilization sound like an actual serious persons proposal that is actually needed to drive player engagement and a sense of risk? REALLY? Or. Maybe. Is it a sign you are putting a fucking LOT of mileage into those terms and concepts?

[Thaluikhain]

If the goal is to present death as a threat without having it actually happen much at all

At the risk of further derailing the thread, that seems like a circle people have tried to square for ages, without a huge amount of success. Making death seem like it could really happen would probably only really work if it could really happen.

(Though, generalisation, but might be best to kill characters at low level before players get attached, and then hope they are still scared of dying at higher levels where you don't really want them to)

OTOH, comic book heroes nominally face death every issue, but the audience knows they aren't actually going to die and are ok with that.

[tussock]

DnD solved the "you might die" problem right from the start, by giving PCs stacks of hit points, and having monsters do not enough damage to kill them in one hit, but that in turn reduced their hit points until one hit would eventually kill them. But you could see it coming, and do something about it before then, if you were paying attention and shit.

Saving throws in the early game were for escaping the things which bypassed your heroic ability to not die to trivial things doing d6 damage per hit, like, "no, that just kills you, but, make a save", and that worked pretty well through to 3e where they (mostly) made monsters use the player's spells and attack rules, including critical hits for 4x damage, with monstrous strength scores and so on, and so on, after spending 25 years giving the PCs ever more lethal spells and attacks. So, suddenly, players faced a lot more of save-or-suck, save-or-die type mechanics, which were fine for Mages to dish out with their severely limited spell slots, and not so much with monsters getting them 1/round (or worse) forever.

They also nerfed Fighters, presumably to make the monsters work less well, but that's an aside. 3e was very good, and also not very good.

Anyway, yeah. If you really want "I can't die yet, but maybe I'll die in the boss fight", then you want stacks of hit points, and limited healing, and relatively low damage per monster, where fights seem less challenging but wear you down a bit, and then later maybe you take a risk on one more fight, because the XP and/or gp system rewards that multi-fight adventure risk somehow.

Death Flag is more if you're playing 3e and want to not die to random crap they gave to some stupid mid-tier monster, or a wild crit from an orc wielding a pick axe or something before you got to 8th level, without house ruling everything in sight. You don't have to eliminate save-or-die, it works quite well for PCs killing chaff monsters after all, saves time, just the bit where the monsters have too much of it, let the PCs ignore that part.

[The Adventurer's Almanac]

At this point, why not give let player characters and boss monsters either auto-succeed on a saving throw or nullify a critical hit through expending a meta resource? Like some kind of... fate point.
But if they spend that and still fuck up, then they should just die. Sometimes players make mistakes and lose the game. It's actually fine. They can play again.

[tussock]

I think it probably makes a better D&D game to define leaders in general, where all PCs are (at least 4th level + or so), and NPCs that are obviously in charge of a group, and like Beholders or whatever.

Where leaders don't roll for morale failures, are immune to social skills, never die (or take more than half max HP in damage) because of a failed save, and all that sort of thing. So you can use the good tropes in the game, like, Finger of Death that is actually point at and cause death, or Charm Person that actually charms a person, but it doesn't kill the fun for PCs to get hit with it.

And so your morale rules would be inherently based off the leader's Charisma, and when you gank the leader the rest tend to fail badly on defaults and run off or surrender, but you can't one-shot him with some stupid spell so it's still a game.

Which, classic D&D almost always did through the numbers it used for saves and stuff, but now that's gone and saves often don't work for anyone, well, just do it all the time instead of most of the time.

[Neo Phonelobster Prime]

At this point, why not give let player characters and boss monsters either auto-succeed on a saving throw or nullify a critical hit through expending a meta resource? Like some kind of... fate point.

Because that option is actually far softer in the elitist hard core pro player angle than the arbitrary post facto "that didn't kill me" option.

Because your fate point option prevents the defeat from happening in the short term. Meanwhile the arbitrary choice over whether a defeat kills a character is only a decision that happens after the defeat and only effects the long term consequences, not the short term consequence of being removed from participation in the current encounter.

Essentially there is a disconnect between how hard core your game play is, and how hardcore your disregard for long term coherent narrative is. Your option does nothing to save the story from suddenly ending because you CAN still run out of video game spare lives (the super gentle automatic temporary invulnerability with no respawn or other consequences type) but ALSO makes the actual game play consequences less impactful. It is frankly, a lose lose option.

Where leaders... ...never die (or take more than half max HP in damage) ... and all that sort of thing.

Oh FFS.

Poorly defined Leaders? PCs being leaders, but only maybe sometimes. Having all the immunity to everything and never dying, never even taking more than half max HP in damage? Randomly deciding beholders should have all this too. Rambling about morale rules?

And your generous fellatio of "old school D&D" (name and advocate for the specific edition COWARD) for the exact opposite of what it is generally known for is sickening and bizarre.

Especially considering what you are saying you actually want to do with total player leader immunity to half the mechanics in the game and anything that might threaten their 4E style sumo hit point sponge longevity.

Death flags was a bad idea but at least it was internally coherent. This? WTF?

[Thaluikhain]

Where leaders don't roll for morale failures, are immune to social skills, never die (or take more than half max HP in damage) because of a failed save, and all that sort of thing. So you can use the good tropes in the game, like, Finger of Death that is actually point at and cause death, or Charm Person that actually charms a person, but it doesn't kill the fun for PCs to get hit with it.

Doesn't that mean that most of the cool and interesting powers only work (or work properly) on chaff monsters that aren't cool and interesting?

Which, ok, that might be what game X needs, but if the game involves you dealing with chaff monsters easily without all that mucking around, what's the point.

Though, instead of marking some people as leaders and some not, I think it'd be easier to just have powers work differently (or perhaps otherwise not work at all) when there's a big gap in levels. Nobody is inherently important, it's just relative. Again, assuming that's what you want to do in the first place.

[Krusk]

At this point, why not give let player characters and boss monsters either auto-succeed on a saving throw or nullify a critical hit through expending a meta resource? Like some kind of... fate point.
But if they spend that and still fuck up, then they should just die. Sometimes players make mistakes and lose the game. It's actually fine. They can play again.

5e did this with legendary resistance for boss monsters. Usually something like 3/day you can choose to succeed on a save that you failed. I actually kind of hate what it did in practice. Because people walk into a fight with something that looks like a boss, and don't throw their big spells at it, instead engaging in some sort of weird feint minigame with the DM. Trying to get the monster to burn their 3 auto-saves on spells that aren't the game enders before finally letting their big one drop.

[tussock]

Heh, yeah, it'd have to be always on, really. Some people just need to take a beating, and whatever the partial effect is on "if you were immune to death" on the finger of death can contribute to that, so people don't feel too bad about trying things. Like it does when you zap a Vampire with it.

Like, there's various monsters immune to those sort of effects in part, and the spells can have some basic damage effects ready in those cases for where leaders are also immune to insta-kill type stuff.

I mean, the problem in 4e/5e for not having insta-kills vs the boss monsters is they gave the big ones a thousand hit points, and had characters doing 20 damage a round at them, and in general you would not want to do that because it's awful, and applies to dragons at least in 3e as well. You do need some way to win, the game should work, rather than not working a la high level 4e vs solos. Though the first step to a functional game in 4e is half the monster hit points and double their basic damage, you know, it's a low bar, but you gotta cross it.

[MGuy]

For killing PCs I decided to go with the meta currency option. Since the currency is used to for more than not dying (it's also used to allow players some limited narrative control over the game). Personally I'm ok with a group just getting wiped from time to time but in practice it sucks a lot of energy out of my players. Using meta currency seems to get me the vibes that ultimately allowing straight up death would give me without upsetting the people I game with. The vibe being that players don't treat dying as a minor thing but if things just don't go their way they have a way to buy themselves out of a jam. Since I've been running just lengthy pathfinder games for the last few years and I find my life too exhausting to seemingly ever get the collection of notes I call a project into a workable state, it's not very formalized yet. In my private work however I've had some version of a meta currency system for years now. I'm sure I made a post about it years ago but I can't remember what I named it. In the state it is in now, ideally, it is going to run off a "power of love/connections" concept where this meta currency is generated by creating and maintaining positive relationships with NPCs, advancing organizations (the player's or NPC based), having a family, etc. The idea being to tie narrative meta currency with increasing narrative relevance. The cost of getting a death 'buy out' will probably grow with level so that dying is never a trivialized in cost as the number connections players have to draw upon grows over time.

I think this works for me because the only way to grind more points is to actually engage with the setting which is my primary goal even beyond leaving an option for killing the PCs open (if very unlikely) inside the ruleset. The tricky part I have yet to work out is how exactly I'm going to codify all those meta currency generating things in the game in a way that 'feels' right. I ran the Iron Fang AP about a year ago now and more recently I've been playing in another friend's cyberpunk red game. I have a few notes I have been scribbling to myself, noting what seemed to get players engaged with the NPCs they are tasked with toting around in it. But that's a different discussion.

As for other characters in game, in general, I'm fine with NPCs dying and an in support of the concept of 'boss ' (or different tier opposition). This really is only relevant to the discussion about PC death if you intend to have 'bosses' be the route you take to introduce death to PCs. I'm not exactly sure what's being put on the table at this point or if the conversation is shifting toward how fast should players be able to put down foes that are intended to be more challenging.

[Thaluikhain]

For killing PCs I decided to go with the meta currency option. Since the currency is used to for more than not dying (it's also used to allow players some limited narrative control over the game). Personally I'm ok with a group just getting wiped from time to time but in practice it sucks a lot of energy out of my players. Using meta currency seems to get me the vibes that ultimately allowing straight up death would give me without upsetting the people I game with. The vibe being that players don't treat dying as a minor thing but if things just don't go their way they have a way to buy themselves out of a jam. Since I've been running just lengthy pathfinder games for the last few years and I find my life too exhausting to seemingly ever get the collection of notes I call a project into a workable state, it's not very formalized yet. In my private work however I've had some version of a meta currency system for years now. I'm sure I made a post about it years ago but I can't remember what I named it. In the state it is in now, ideally, it is going to run off a "power of love/connections" concept where this meta currency is generated by creating and maintaining positive relationships with NPCs, advancing organizations (the player's or NPC based), having a family, etc. The idea being to tie narrative meta currency with increasing narrative relevance. The cost of getting a death 'buy out' will probably grow with level so that dying is never a trivialized in cost as the number connections players have to draw upon grows over time.

I think this works for me because the only way to grind more points is to actually engage with the setting which is my primary goal even beyond leaving an option for killing the PCs open (if very unlikely) inside the ruleset.

Oh, you could easily use that as divine patronage or something you get from doing quests and stuff in a fantasy setting, definite possibilities here.

[merxa]

so MGuys implementation isn't too far from my prior suggestion of tying narrative death to narrative powers, instead its attempted to be made fungible with a meta currency (where you pay x amount of tokens to avoid death and perhaps x - y tokens to succeed at some task that has narrative implications).

MGuy may want to investigate the pulp cthulhu luck mechanic to see how one game handles such a meta currency.

But i am curious how PCs earn the meta currency, it seems like they gain it via establishing in game relationships to npcs? MGuy mentions it is earned via 'creating and maintaining' positive relationships, so it is earned on the initial relationship being established then there's a reoccurring meta currency generation, but that may require some level of in-game relationship maintenance to ensure? I could see how formalizing some of this could be hard as you may not want to gamify standard roleplaying, but still encourage players not be murder hobos.

one thought is to tie it to a domains minigame, where you have vassals who pledge support, and then additional sources from having a family, maintaining trusted friendships with important NPCs, running key organizations. Presumably you would want some sources to also act as sinks, so perhaps running a kingdom has a maintenance cost, influencing other organizations, diplomacy with other kingdoms etc.

The primary issue I have found with providing a metacurrency to avoid death, is players will naturally horde the metacurrency for such an event, and if they ever are below the amount needed, there is a strong tendency to become extremely risk adverse and do whatever they must to get their meta currency back above the threshold. I think it tends to create a little too much meta gaming and at extremes warps the narrative with PCs taking actions that wouldn't make sense if the meta currency wasn't present.

[The Adventurer's Almanac]

Personally, I think the best option is to just start killing player characters. You have to fuck up really hard to die in something like 5e. The only time I've seen it happen is when enemies go for the kill after the PC is down, which is generally understood to be unsportsmanlike behavior anyway - or done for narrative purposes. Do you guys play with people who get mad about eating shit and dying all the time or something?

[Kaelik]

To give away a few of the DM tricks to talk about design a bit.

I think it's probably fine to just kill characters mostly, but you should give players of higher level more narratively difficult to replace characters a thing they can choose to do which allows them to escape death that at least theorteically can be defeated by the opposition, and then let the players self select how much they want to protect their characters.

IE, Wizards can cast Clone, or can use Lesser Planar Bound Nightmares to astrally project, and become immune to most deaths, but theoretically they can be undone.

Then, when a player starts getting SUPER paranoid about where there body is, probably don't kill that player character, but do send some fun attacks, or kill their Astral Projection more often, so they can feel cool about having beaten the rap. But if someone COULD cast Clone, and just doesn't, then don't kill them as often as the Astral Projection, but you don't have to hold your punches, because that's a player who is totally fine having their character die.

But for 3e you do have to first change things somehow so everyone has reasonable access to stuff like this.

For example, my Kaelik Barbarian has an item that creates a temporary clone of him with less HP and has that adventure, so the DM should feel fine killing that in fights all the time, but obviously I am trying real hard to stay alive. On the other hand Radthemad4 has a character which summons and calls a bunch of outsiders, but never uses Astral Projection or avoids combat, so presumably would be fine dying, but the DM should definitely not go out of his way to kill rad's character like he should to kill my Decoy.

[Foxwarrior]

Ooh, a narrative/feelings based answer, secrets are revealed indeed! So you see, in Kaelik's games fighters are OP and cannot die, because they have no features that protect them from death which they can neglect to use...

[gatorized2]

Save or X doesn't have to be instant, and it doesn't have to bypass defenses, or have its own special set of dedicated defenses. We can also provide additional ways for the players to interact with anti fun effects so that they don't feel like bullshit.

Personally, I avoid using such effects against the players, while allowing the players to use them as much as they want. This has made my games way more fun and easy to run, for myself as well the players.