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

[Seerow]

If you want your liches to snuff out the lifes of peasants like candles hand them a spell that does xd6 times 10 damage if this is enough to kill the target, xd6 times 3 otherwise. Flavor of death magic conserved, stacks with "real" damage, can be used against major characters and peasants alike.

Can you go into detail as to how this would work, or why you would wnat to do it that way?

It sounds like you're saying you give an attack that deals say... 10d6 damage. If 10d6 rolls say 40 damage and kills the target, they take 400 damage. If it rolls instead 30 damage, just shy of killing the target, they take 90 damage, which then kills the guy who would have died with 40 damage anyway.

What's the point of the multiplier? Why give a stronger multiplier if it kills the target? Why give a multiplier at all if it doesn't? Your example either makes no sense or I am missing something obvious.

[Murtak]

I got no problem with it for Slay Living, or even Sleep (which could just deal nonlethal damage). But trying to make Charm Monster do HP damage gets all stupid so we need another solution anyway. We might as well figure out that solution first just in case it also works better for Slay Living.

You could go with a separate damage track for things that attack your mind - sleep, fear and charm effects would fit well on the same track in my opinion. You need to take care to not create too many damage tracks or you are right back with damage effects not stacking, but just hitpoints and willpower should work out decently, especially if there is the odd power that allows conversion between tracks or that allows you to take damage to one of the tracks for effects that thematically belong to the other track (i.e. take damage to regain willpower, spending willpower to fuel power attacks, etc.).

[RadiantPhoenix]

What's the point of the multiplier? Why give a stronger multiplier if it kills the target? Why give a multiplier at all if it doesn't? Your example either makes no sense or I am missing something obvious.

My interpretation was thus: roll Xd6; if the result times 10 would kill the target it does that, otherwise it deals the result times 3.

[Seerow]

What's the point of the multiplier? Why give a stronger multiplier if it kills the target? Why give a multiplier at all if it doesn't? Your example either makes no sense or I am missing something obvious.

My interpretation was thus: roll Xd6; if the result times 10 would kill the target it does that, otherwise it deals the result times 3.

Ah that makes more sense, as long as X remains relatively small.

[Dogbert]

"A balanced dnd system" is an oxymoron, just wanted to point that out.

Also, the thing with rules is that what's good for the goose is good for the gander, and Rocket Launcher Tag mechanics got removed from game because Mister Cavern tended to kick and scream every time he got his RPeen NPC out to cockslap the players only to get 1-shot by a lucky wizard.

[OgreBattle]

"Fate" mechanics to survive death X amount of times or risk it is the most straightforward way to do it.

[Thaluikhain]

I like the CAN thing linked to.

I'm also reminded of the vehicle damage table for 40k in some 3-7 ed, can't remember which. Roll a d6, add modifiers, and you need a 7 or higher to instant kill. So normally you'll just damage the thing, and if it's been damaged before you might kill it this time, but with luck and some bonuses you can one shot it.

(Also, nice the spambots necro-ing interesting posts)

[Kaelik]

I like the CAN thing linked to.

I'm also reminded of the vehicle damage table for 40k in some 3-7 ed, can't remember which. Roll a d6, add modifiers, and you need a 7 or higher to instant kill. So normally you'll just damage the thing, and if it's been damaged before you might kill it this time, but with luck and some bonuses you can one shot it.

(Also, nice the spambots necro-ing interesting posts)

Fiends and Fortresses has condition tracks for everything. I think CAN is supposed to be a single condition track, but I took a lot of influence from that.

[Grek]

Something like the Power Word series of spells or Murtak's lich touch suggestion is best.

I'm currently tinkering with rules where you roll Xd6 to attack, and if the sum exceeds the target's hit points, they just die. Otherwise, they lose hit points equal to the single best die rolled. For moves that are simply save or suck rather than save or die, you take 2x the best die in damage on a failed attack roll, but the defender has the option to eat the damage instead of the status effect if they think that would be better for them.

[Bigdy McKen]

Maybe I’ve been brain poisoned by too much Elden Ring, but you could just use condition tracks that build up over time, similar to the way that death, madness, frostbite, etc. work in that game.

You wouldn’t even need to make a discrete track for every single unique condition, just make general tracks for fortitude, reflex, will and maybe awareness for illusion/trickery type stuff. Each spell can then have an additional, unique effect like a debuff that happens as long as the spell is in effect.

I’m thinking spell caster hurls a petrify at the target who makes a saving throw each round, while caster spends an action to keep the spell going. Each failed save moves them one step toward the bad end, and each successful one moves them one step toward the good end.

Might work better within a 5E paradigm since you’re only getting one slot at the levels where power word kill and finger of death come online rather than 3E where you get upwards of 4 slots for each spell level.

[deaddmwalking]

Maybe I’ve been brain poisoned by too much Elden Ring, but you could just use condition tracks that build up over time, similar to the way that death, madness, frostbite, etc. work in that game.

You wouldn’t even need to make a discrete track for every single unique condition, just make general tracks for fortitude, reflex, will and maybe awareness for illusion/trickery type stuff. Each spell can then have an additional, unique effect like a debuff that happens as long as the spell is in effect.

I’m thinking spell caster hurls a petrify at the target who makes a saving throw each round, while caster spends an action to keep the spell going. Each failed save moves them one step toward the bad end, and each successful one moves them one step toward the good end.

Might work better within a 5E paradigm since you’re only getting one slot at the levels where power word kill and finger of death come online rather than 3E where you get upwards of 4 slots for each spell level.

The devil is in the details.

If I'm a caster and I cast petrify, am I losing the ability to contribute anything else other than maintaining my spell? Does my target have to spend an action resisting the spell if they make their save? A spell that doesn't do anything to the target (because they make their save) but effectively paralyzes me (because I have to spend my full turn maintaining the spell) is probably not a solution you want. And if the track requires 5+ saves to be effective, it's going to struggle to be combat relevant.

Before you ask 'how can we do this', you should at least take a stab at answering 'should we do this?'.

While Save-or-Die has a history in the game, from a design perspective you should consider what it gives you and whether that's even what you want. Is it 'good' that a world-ending threat has a 5% chance to be obliterated, or would that type of victory be anti-climatic. I think it's defensible to say 'you shouldn't be able to one-shot level-appropriate challenges', without taking away the ability to do so to lower-level opposition. And of course, depending on what you decide you want, there are multiple different ways to approach the problem.

Having multiple (non-fungible) tracks can be a problem. If Character A is attacking the Reflex Track and Character B is attacking the Fortitude Track, but damage between them is completely separate, effectively one of the characters is wasting their time. One of the advantages of hit-point-damage is that it is easy to make sure everyone's contribution matters. If an opponent has 100 hit points and you deal 20 damage to them, you've contributed something even if another character does 80 damage and finishes him off. On the other hand, if you do 20 damage and then someone else banishes that opponent, your damage didn't contribute at all.

Adding de-buffs to more attacks may help to some degree. If a Reflex Attack makes the opponent easier to hit, allowing the Fortitude Attacks better chance of succeeding, you could make the case that each character's actions are contributing meaningfully to the defeat of the opponent, even if eventually only one track involves the 'final blow'.

And if you consider putting anything on a track, you have to consider how people 'heal' that damage. Obviously players are going to want to start every fight they can at 'full power', so they won't have much incentive to start a fight with 50% fortitude damage. In a video game it's pretty easy to establish a scenario where a track will persist until the scenario is concluded and that can create both an incentive to keep moving (complete the scenario as quickly as possible, before the damage accumulates to lethal levels) and organically creates additional drama regarding 'end-boss' encounters where you are going to start the fight depleted of resources and the question is really how weak are you as a result of the time/actions you took to that point.

It's easy to describe how you think something might work in one single 'clear' case, and that's a great starting point, but once you look at implementing it you're going to have dozens and dozens of other questions. Going back to the example you provided, if I cast petrify and someone else also casts petrify on the same target, what happens? Do they potentially petrify 2x as fast? Do they make one save but the TN is more difficult? As you answer those questions you're going to create strategies and synergies that you might not have intended. As you discover them, you definitely need to consider whether they generate the outcomes you want - which is really the whole point of the game rules in the first place.

[Thaluikhain]

Hmmm...what about instead of petrifying them per se, you have a spell that does cumulative damage to, say, their Reflexes, and if it hits 0 the target dies (is petrified for flavour), if it doesn't get that low they just are slower and so do worse in combat? So then attacking their Fortitude would also be useful, in that it'd weaken them in another way, even if it doesn't stack with getting their Reflexes to 0.

Though, might be a book-keeping issue with fiddly details depending on the system.

[Kaelik]

Though, might be a book-keeping issue with fiddly details depending on the system.

Fiends and Fortresses has many condition tracks that contribute at crosses to each other and..... while it is very uncommon for anyone to be affected by more then three things at time so far, it is really clear that an automated character sheet that tracks the cross penalties is while not strictly mandatory, very very very very useful.

[OgreBattle]

I think Save or Dies should be approached in regards of D&D being a team based game

So... going in alone into the Tomb of Bulldookie you walk into the portal and die
But going with friends there's an opportunity for someone to also roll a save to save you

This is just a formative idea, but you know there's all kinds of action stories where someone saves someone else like tackling out of the way or halting them before triggering a trap or knocking a fatal arrow out of the air.

I once saw a little girl pull her mother crossing the street out of the way of a speeding construction equipment laden truck, in D&D terms she'd need class levels in Aegis Defender Knight to do this, but that kind of thing happens in 'reality all the time'

[deaddmwalking]

So... going in alone into the Tomb of Bulldookie you walk into the portal and die
But going with friends there's an opportunity for someone to also roll a save to save you

This is just a formative idea, but you know there's all kinds of action stories where someone saves someone else like tackling out of the way or halting them before triggering a trap or knocking a fatal arrow out of the air.

From a narrative perspective, an arrow that could have killed you but didn't looks a lot like an arrow that couldn't have killed you. Either way, your character isn't dead after the attack. From a player perspective, needing to be saved, especially from a relatively minor threat, doesn't feel heroic. Additionally, failing your save doesn't guarantee that someone else will make theirs, meaning you can still die... Ie, I step on the floor space that causes a 10-ton block to fall from the ceiling and I fail my reflexive save, but the GM says that the person next to me can push me out of the way with a TN 20 Reflex save and they roll a 15... Well, you have one pancaked character.

Within a single game having an ad-hoc situation where someone has a 5% chance to die, followed by a 95% to negate the roll (0.25% actual chance of death) you're probably going to be okay with the result. But across 400 games, that same situation will create a very unsatisfying outcome. As a designer, you can say 'I can live with the game doing stupid things 1/400th of the time' because getting things to 'usually work' is relatively easy; but if you want to constrain outcomes to generate more satisfying game play setting the odds to be 'relatively low' isn't really satisfying. If you don't think the trap should EVER kill the mid-level character without some additional factor in play, you should consider something other than probability. That could take a lot of forms including a meta 'fate' currency that you can 'burn' to be saved (in which case the character that would have died narrates that his friend pushed him out of the way).

Iterative probability is a bitch. If you have a 1% chance of a TPK in every combat, most groups will experience one before they have 100 separate combats. You want it to FEEL like there is a real chance of danger; like the odds of a TPK are significantly CLOSER to 10-20%, but you probably want to make it much lower than that. Much lower even than 1%. It's okay to make it 0% provided the players haven't chosen to spend whatever meta-currency or abilities you've given that allows them to miraculously survive - sometimes it is nice when you can experience defeat as a party without ending the campaign.

[Omegonthesane]

I appreciate my group is not a representative sample of the world - but when in situations where we are given a meta-currency that can be spent to Not Die, we proceed as if real character death is actually on the table and then just spend the currency when necessary.

[deaddmwalking]

I appreciate my group is not a representative sample of the world - but when in situations where we are given a meta-currency that can be spent to Not Die, we proceed as if real character death is actually on the table and then just spend the currency when necessary.

I'd wager that you're more representative than you realize.

If it's baked into the rules, then it isn't anything 'special' and people would feel stupid/bad for NOT using it. There are people that will go their entire life swearing that something is stupid/bad (like Social Security) but won't refuse to accept it when they qualify. Mostly they just don't want it for OTHER PEOPLE. And if the implementation is relatively straight forward and is genre-appropriate, it won't create disassociation at the table. When character death comes up, most people will desperately look for an escape. If someone falls off a cliff, I guaran*damn*tee you that their next question is 'is there a branch that I can grab on to'?

Rather than GRANT GM Pity (which some people WILL object to), build it in to the rules that there's a way they can miraculously survive whether that's catching a branch on the way down or breaking every bone in their body but somehow still not be dead (and yes, it does happen in real life).

[Kaelik]

I think Save or Dies should be approached in regards of D&D being a team based game

So... going in alone into the Tomb of Bulldookie you walk into the portal and die
But going with friends there's an opportunity for someone to also roll a save to save you

This is just a formative idea, but you know there's all kinds of action stories where someone saves someone else like tackling out of the way or halting them before triggering a trap or knocking a fatal arrow out of the air.

I once saw a little girl pull her mother crossing the street out of the way of a speeding construction equipment laden truck, in D&D terms she'd need class levels in Aegis Defender Knight to do this, but that kind of thing happens in 'reality all the time'

I think you should strongly consider not adding a bunch more rolls to every spell of effect.

[Thaluikhain]

About having Not-Die currency, would that be it's only use, or would you use it for other things? If it's only used for that, you don't have the problem of it being wasted and not available when needed for important Not-Dying, though.

[deaddmwalking]

About having Not-Die currency, would that be it's only use, or would you use it for other things? If it's only used for that, you don't have the problem of it being wasted and not available when needed for important Not-Dying, though.

That's a decision you have to make. Personally, I think having a 'not-die currency' that doesn't do anything else feels like too much of a kludge. Having a resource that does something and keeps you from dying means you have the ability to spend all of it and leave yourself open to death, but most people most of the time will reserve at least one token for that most important function.

In the Heartbreaker I play in, we have 'Action Points' that serve as a 'not-die currency'. You can also spend an Action Point to change a die roll for an attack or save to a '10', spend one as a move-action to take a breather (regenerating a small number of hit points and spell points). There have been a couple of times where players have used their last action point while combat is still happening - it definitely increases the stakes - but that's also a player's choice. You can get 1 Action Point back with 1 hour of rest and most characters have around 6 (current party has 5, 6, 6, 8). Essentially, combat uses up your small number of available resources quickly, and action points can regenerate your combat ability mid-combat, but your ability to continue refreshing throughout the day degrades as you deplete the Action Points. For us, that helps balance the '5-minute adventure day' - if you're full on all your abilities except Action Points your combat effectiveness is 100% at least for 1 more fight; but if it uses up your hit points/special abilities you might be less effective by the end and really worn down.

This also ties to the Combat Ability Number (CAN) conversation to some degree. Since Action Points exist for both the PCs and MAJOR NPCs, if you can succeed against a SOD by rolling a '10' or better, any character/major NPC can avoid the '1-shot' with a poor roll, but it still depletes their combat effectiveness because they're spending a limited resource-regeneration currency. Spamming an attack like 'blindness' that could make a tough enemy very easy to defeat may force them to spend the meta-currency so they're not regaining spells or modifying their attacks, or activating expensive special abilities. A non-major NPC doesn't have Action Points, so you might defeat them if they roll a bad save, and if you faced an important NPC that has Action Points but can't succeed on a save with a 10 or better, you outclass them enough that it is possible to 1-shot them with luck.

So while they do serve as a 'not-die' currency, they have enough other functions that we use them (multiple times) every session and players have ownership of it (as opposed to an exclusively GM-fiat system). For us, it helps maintain the illusion of a truly adversarial GM who is fairly adjudicating the opposition - bad guys don't have to pull their punches when they're winning to avoid a PC death. Perhaps importantly, while this allows you to 'not die', it doesn't allow you to necessarily stay in combat.

Broad lines of the design decisions, we originally had Action Points activating many attacks/abilities directly. We found that players didn't get to activate special abilities enough, and we didn't want to give out so many of them that managing them as a resource wasn't interesting. My character has 7 spell points; it costs 1 to designate an enemy to deal extra damage for the combat; it costs 1 to get an AoO under specific circumstances; it costs 1 to ignore a detrimental status effect for 1 round. This character has 5 AP, and can spend one to regain 9 hit points and 3 mana.

By no means am I suggesting that this is the only or even best solution to this problem - but this happens to work for our group and helps generate additional tactical depth in our combats. Characters are frequently spending an action during combat to regain some power. For wizards, particularly, a lot of their most powerful spells require a full-action, so 'sucking mana' (regaining spell points) encourages them to cast a spell that is less powerful (requiring only a standard action) regularly.

[Omegonthesane]

So, the examples that I know my group has repeatedly flirted with over the years are Destiny Points from SIFRP and Fate Points from Dark Heresy*.

In both instances, you have two broad categories of uses:
* "spend" the points to get certain minor benefits to tip the odds in your favour, like a bonus to your roll or a reroll, and get them back at the end of the session or the end of the adventure; or
* "burn" the points to get a much more powerful benefit, on a menu that includes "cheat death", and you don't get the point back. Ever. You can maybe buy new points with EXP but you are just permanently that bit weaker than if you had survived without burning the point.

* technically we played Adeptus Evangelion, then Rogue Trader, then Black Crusade, with other campaigns in between; and the mechanic is Infamy Points in Black Crusade and has some other convolutions besides; but they're iterations of the same underlying system and I am unaware of any collective term for the shared underlying engine

[deaddmwalking]

If you 'burn' points, bring them back at some point (like when characters gain a level). First off, a character that is built at higher levels (rather than played from 1st level) will never have had to worry about burning those points, so they have an advantage; and secondly, players won't want to burn them even when they should. If you are still carrying around potions in a video game, you know that people don't want to use something if they can't get it back. Recognize the psychology; let them earn it back after some appropriate amount of time.

[merxa]

the more general question is, should PCs die? My answer is yes, it should be on the table, there should be some set of decisions and/or dice rolls that cause PC death. And if we were perfect arbitrators of destiny, what is the expected rate of PC death, every 100 sessions or so on average? What if someone never makes a 'bad' or non-optimal decision, should death now be impossible for them, or does the expectation go to 200, 300, 500 sessions?

In the most extreme circumstances, a PC could avoid all risks and become a pastry chef, most games would probably retire that character and move on. And short of making some epic cooking competitions

it would be assumed by most players to be a GM ass-pull to reign down terror on the pastry chef (the most generous view could be the GM had long planed to nuke the city they happened to setup shop).

Another reason for death to be on the table is why else have raise dead? Sure it can be used on NPCs, but its cooler to use your spells on central characters including PCs. Another argument is decisions, presumably a PC could make a series of decisions that lead to death -- it need not be so self-evident as a PC stating they commit suicide and the GM says no -- but if decisions cannot lead to death it trivializes their own agency and players will notice, eventually.

So what about dice? the standard d20, on a single roll, you're looking at 5% as your smallest probability, which seems pretty high for D&D land, getting confronted with a sudden 5% chance of death from a single roll does feels daunting, and if that happens once per session, after about 14 sessions you would expect half the PCs to be dead. If say rolling 1 twice on a d20 (1 in 400) happened once per session, you would expect half the PCs to be dead by 280 sessions. In my preferred 3d6, rolling 3 (~.4629%), you get 150 sessions.

Most campaigns don't hit 150 sessions (that's about 3 years at 1/week). But with a 4 player table, you might see a death after about 35, 40 sessions.

To me, that doesn't seem so bad, especially when you consider the fact that being dead-dead in d&d land gets increasingly difficult as you level. So if you're using a d20, I'd probably use a condition tracker, so petrification would go from normal - staggered - petrified, maybe include one more step at most. If you only have one or two steps before the end condition, its easier to make those inbetween steps a significant debuff, so even if everyone else is dealing hp damage, your ray of flesh to stone could still be helping out.

[Kaelik]

I think you should generally just design your game to output death a lot less, and then not have raise the dead spells or narrative currencies to undie.

If the game is designed with a lot of Flesh to Stone or Sleep effects instead, but the party can get them back if they don't TPK, that's probably just better. You do want a certain amount of death possibility, but the intention really should be that an enemy almost never drops someone from active to dead, and if they are dropped, that there's little incentive for enemies to kill them post incapacitation.

This is if you are redesigning. If you are just paying D&D, you should probably just let people die as much as D&D does, and if they don't want to die very badly let them get Deathward, and probably houserule death at -10 HP to some much larger number, but otherwise, just have everyone use a staff of Life or make a new character or whatever.

[deaddmwalking]

the more general question is, should PCs die? My answer is yes, it should be on the table, there should be some set of decisions and/or dice rolls that cause PC death.

I agree that PCs should be able to die, in part because that's an expected part of genre simulation. Trivial PC death (such as stepping on a trip wire - rocks fall - you die) is bad, and exists in traditional D&D. There's a larger issue with traps generally - for traps to be FUN they must be engaging. Essentially, that means you have to know you're in a trap and have a chance to interact with the trap and/or the mechanism. But that's tangential to save or die generally.

So what about dice? the standard d20, on a single roll, you're looking at 5% as your smallest probability, which seems pretty high for D&D land, getting confronted with a sudden 5% chance of death from a single roll does feels daunting, and if that happens once per session, after about 14 sessions you would expect half the PCs to be dead. If say rolling 1 twice on a d20 (1 in 400) happened once per session, you would expect half the PCs to be dead by 280 sessions. In my preferred 3d6, rolling 3 (~.4629%), you get 150 sessions.

Most campaigns don't hit 150 sessions (that's about 3 years at 1/week). But with a 4 player table, you might see a death after about 35, 40 sessions.

Keep in mind that these are non-cumulative chances. If something happens 1% of the time, we can say 'it'll probably happen approximately once every 100 times', and that's true. But it is EQUALLY LIKELY to be the 1st time or the 100th time. And in fact, it can be both the first and second time and while that's unlikely, it isn't actually surprising. Something that happens 1/10,000 times WILL HAPPEN IN SOME GROUPS if you have enough groups playing. And as you point out, having multiple players each with the same chance multiplies those odds.

Think of it this way: what do the odds have to be before you feel comfortable playing a game of Russian Roulette? What if you were expected to play that game hundreds of times?

To me, that doesn't seem so bad, especially when you consider the fact that being dead-dead in d&d land gets increasingly difficult as you level. So if you're using a d20, I'd probably use a condition tracker, so petrification would go from normal - staggered - petrified, maybe include one more step at most. If you only have one or two steps before the end condition, its easier to make those inbetween steps a significant debuff, so even if everyone else is dealing hp damage, your ray of flesh to stone could still be helping out.

D&D is a cooperative game. If being dead means you can't participate in the game for some number of real-life hours or multiple real-life sessions that's not ideal. If death is a temporary, easily removed condition (like hit point loss is in standard D&D), I don't think that gives death the seriousness that it deserves. For what it's worth, we don't have any [real] way to bring a dead character back to life. We also don't have to have 'super-dead' like 'disintegrated' or 'eradicated'.

Saying 'death can happen but it is super-easy to remove' is virtually the same as saying 'death can't happen', but it is the worst version of that to my personal taste.