Players don’t just want to win. They want to win by the skin of their teeth. They want to feel like the dragon could have eaten them and they just barely got away. So how do you create that feeling without actually killing any characters? That’s what we’re talking about this week, including everything from how to make your villains threatening to what happens if you actually do kill off PCs whenever the dice say to. Also, a few strolls down memory lane to the campaigns of yesteryear!
Transcript
Generously transcribed by Olivia SB. Volunteer to transcribe a podcast.
Chris: You’re listening to the Mythcreants podcast, with your hosts Oren Ashkenazi, Wes Matlock, and Chris Winkle.
Oren: And welcome everyone to another episode of the Mythcreants podcast, I’m Oren, with me today is:
Chris: Chris
Oren: and
Wes: Wes.
Oren: All right, I’ve rolled criticals on my orc attacks, so both your characters are dead. Sorry, that’s just the way it is. I play hardball with you, Death, and, oh, but wait, wait, maybe not, maybe there was a secret magic spell that got cast on you so now you’re alive. Wooh!
Wes: What?
Chris: I don’t know, I feel like death was much more tragic and emotionally satisfying. Don’t you think that spell was kind of a deus ex machina, really?
Wes: Yeah. I’d already started making a new character sheet, I’m over that character, Oren.
Oren: All right, so today we’re talking about the illusion of death in role-playing games. This topic came to me, it was actually very fascinating, I was scanning around the role-playing internet, like I do, and I found an article where some folks had basically decided to model the cave troll fight in Lord of the Rings using the Torchbearer combat system.
Wes: Interesting.
Oren: It was like, okay, cool, this is a fun way to explain the way Torchbearer combat works to a new player. That’s reasonable. But as I looked at it, I noticed that they had to really manipulate the die roles to get the cinematic narrative results that they wanted. Basically the orcs had to roll like crap to avoid killing any of the characters, because we couldn’t actually have any of the characters die in this fight, right, because we need them later. And so it’s like, okay, so you’re telling me that in these situations where the orcs clearly had an advantage, they all just happened to completely whiff their die roles? It’s like, ‘Um, hmm, er’. I don’t buy it.
Chris: Yeah, it’s hard, because in that fight, it’s clear that the troll and the orcs just have higher stats than the hobbits. If they didn’t it would be silly, there would be no threat to the fight, but at the same time, somehow they lose?
Wes: They’re just fated to fail, maybe?
Chris: And in a D&D style type combat, that randomness of, ‘Well, they had the advantage, but they failed anyway’, is only represented through bad dice rolls. But that doesn’t really- the game is also not set up to just have- it doesn’t work, that’s not how it’s supposed to run.
Oren: Right, and so before we get into it any further, I just figured it would be good to talk about why this is even a thing. Like why not just kill characters when the dice say to and not worry about it?
Chris: I mean, I also think it’s worth thinking about why did we even ever think that was a good idea in the first place?
Oren: Right, okay, well I can tell you that: the reason why we thought that was a good idea in the first place is that role-playing games evolved out of war gaming. And so that was just, yeah, I mean, you rolled good and something died or you rolled bad and you died. That was how war gaming worked. No one tries to impose a narrative on chess, right, it just happens. So that just was the natural conclusion. But then people got weird about it and started putting story into their role-playing game, and it’s all touchy feely now.
Wes: Gas.
Oren: Yeah, I know right?
Chris: So yeah, I’ve got to say that definitely there does seem to be a trend of RPGs getting more into story, although they don’t always know how to do that.
Oren: No, but they want to.
Chris: Yeah. But first of all, stories are generally better with some level of planning, or revision, but of course in role-playing there is no revision. We only do it once, right? So the only option that we have for making the story fit together better is planning. And obviously you can’t plan if you don’t know who’s going to die. It’s very difficult.
Oren: Yeah, right, well that’s what happens when you just randomly kill characters. There’s a couple of things. First, that just ruins a player’s day. If the player was invested in this story and seeing what happened with this character it’s like, ‘Whatever happened to my masked nemesis that I both hated but was also strangely attracted to?’ It’s like, what’s going to, where’s that going to go? And it’s like, ‘Well, I critically failed my role to avoid traps, so that’s going to go nowhere.’
Chris: Yeah.
Oren: And it’s like, ‘I got invested into this character. I wrote backstory. I put on an accent for this character and now they’re just gone?’ And it’s like, well, that sucks.
Chris: Yeah, and conversely, people don’t want to emotionally invest in a story if they think that the characters they are emotionally investing in could randomly die in a completely unsatisfactory manner. And this happens not just with role-playing games, it also happens with shows, where some shows- Lost got to this point for me- where they get to a point when they have killed characters and replaced them so often that it no longer feels worth caring about the new characters they introduce, because you feel like the chances of them dying are just so high that, you know, and then once you lose investment, why watch it anymore, right? And so that fact happens with too much random death- and it’s not just death, but the fact that if you randomly die, you can’t plan any kind of arc around that death or any way to make it satisfying. So then the players have no reason to emotionally invest in the story, which just doesn’t work for stories. You just can’t do a compelling story if people won’t emotionally invest in it.
Oren: Yeah, and I mean the more characters you kill, even if the players are trying to invest in it, even if they are willing to be good sports and don’t just check out of the process entirely, you eventually get to this weird place- and I’ve actually run campaigns where enough characters died that this happened- where you realize that all the people who started the story, who had the original reasons for going on this adventure, are gone. And now it’s just their replacements and you’re like, ‘Why are we all still here guys? Does anyone here have a reason to want to stay on this quest?’ And everyone’s like, ‘I guess not.’ And then it gets weird, it’s just really- the game starts to get listless and everyone’s like, ‘Yeah, I guess we should beat that boss because, you know, that boss killed my brother, except wait, no, he didn’t. He killed the brother of the character who was in the party before me.’ So … eh?
Wes: Yeah, I ran the adventure one time and I killed everyone. It was the third session.
Chris: Oh no.
Wes: And it was interesting because I was very optimistic that they were going to get out of it. And it wouldn’t have been hard to run away, because if you’re familiar with D&D monsters, it was a roper, which can’t really move. I think it can, but not that fast, it’s like a giant, stalagmite? Which one comes out of the ground?
Oren: A pointy thing, you know.
Wes: Yeah, the one that comes out of the-
Chris: Yeah, no you’re right, the stalagmite comes out of the ground.
Wes: Comes out of the ground, yeah. And so like one player drops and I was like, uh oh, are they going to run away yet? And then like another player drops. And I was like, they’re definitely gonna run away now, right? Then they were all dead except for one, and I was like, uh, okay. So we just ended the session, and I was like, okay, well they died and I need to write a little recap email. So I just narrated the last character’s death as well, with his consultation a bit, and then we just called it a close.
Chris: Oh no, that’s terrible!
Wes: And it was sad. We often talk about it when we play, because on the one hand that made death very real for them, in adventures since then they’ve wisened up a bit, maybe- like there are other options instead of just standing there hacking away at something with an insane armor class- and then on the other though, they’d invested in those characters, they still talk about liking those characters and wondering what would have happened with them, and I had planned out a lot more of that adventure as well. It’ll sit around for a rainy day I guess.
Chris: Yeah. I mean, it’s also worth mentioning that a lot of times players don’t seem to know that they can retreat or that they should.
Wes: Yes, please.
Chris: Because they think that the GM expects them to continue fighting, you know, etcetera, etcetera.
Wes: Right.
Oren: Right, I mean, well, it should be pointed out, there’s a whole lot of things going on here. One problem is that yes, most players don’t think of retreating if you’re playing in a traditional D&D style combat game. Because retreat’s not really part of the mechanics- theoretically you can do it, but none of the mechanics really reference it in any way.
Wes: True.
Oren: And against most monsters it’s actually pointless because- there are some, like the roper, that just have a low movement speed, but most of them can move at least as fast as the character. And so if you try to retreat they’ll just chase you. And so players just don’t think of retreat, and then narratively they don’t think of retreat either, right? Because they’re used to- in a narrative, when the fight gets really tough, that means you stick it out and win and it was even cooler because of how tough it was. Not, ‘Oh, well we’ve lost 60% of our tactical assault force, this is clearly a doomed mission. Let’s call for Evac,’ right? We’re not planning this in CQC terms, okay? We’re not commandos. So there’s a whole set of problems there. And then there’s also- this is another issue. There are definitely some players who, if they think you won’t kill their character will dare you to. I’ve definitely noticed this, this is a problem, some players, if they think that you’re being too easy on them, either because of something you said or because of something they see at the session, they’ll just perform more and more outrageous things that really should get them killed.
The solution to that is different for every character. Sometimes you just do have to kill their character and be like, ‘Okay, you were literally asking me to do that.’ Sometimes you can solve it by talking to them aside, that depends on your group dynamics. I would point out that- and I get this response sometimes when I’m like, ‘This sort of thing isn’t really very fun’, and people will be like, ‘Oh, but we still tell stories about the time that happened.’ And I’m like, ‘Okay, I’m just going to say this. Having a bad thing happen so that you can tell stories about it later is not a great way to live your life. You know, I still tell stories about the time that I fell and got a really, really bad cut on my arm that I had to go to the hospital for, but I don’t try to fall and cut my arm so I have a cool conversation piece. That is not a thing, it’s just not a healthy way to live. So that’s just something to keep in mind.
Wes: And the tough thing is that we just still have to dangle that reality that they- we have to, as Game Masters, have to somehow still persuade them to actually be cautious, to have such attachment that they do feel that their character’s in danger, right? It’s like, your ability to run the game and their ability to trust in your running of the game, and as well as embrace the story, means that accepting the reality of death, even if it’s not going to happen, has to kind of be a big component of that. And that’s really hard to do.
Chris: Right. Or there has to be some stakes, right? I don’t know if it would necessarily always have to be death as long as there’s consequences when they fight, there’s consequences of a failed roll. I do think sometimes it’s helpful to have death on the line, even if really the players know it’s not going to happen, if it still feels like in the story scenario it could happen. I mean honestly we do this with non-interactive stories all the time. Most of the time when we go to a movie we know that the protagonist is not going to die in that movie, and the tension is definitely higher if you really think the protagonist could die, but we don’t actually always need the tension to be quite that high, right? It’s okay to have a compelling story with lower tension if the player is emotionally invested enough, then lower tension works better, I would say.
Wes: So you have to Song of Ice and Fire it. You need to kill someone, an NPC that’s beloved, real early on and then not kill anyone else ever again.
Chris: If you want to convince your players- well, an NPC, yes. I mean, if you want to convince your players that they’re going to die when they’re not going to die, then I suppose you could have a mole in your player group who’s like, ‘Hey, by the way, your character is going to die, but pretend this was random.’
Wes: I mean that might be fun.
Oren: Yeah, so from a GM perspective there’s a lot of tap dancing that you can do, because you’re right, this is why we call it the illusion that death was a possibility, because victory means more if you feel like you could have failed. And this isn’t even just a death thing, like in certain situations this can apply to any kind of failure. Like, you get to the end of the campaign and the GM has set the stakes, and it’s like, ‘Yeah, the evil Elven empire is coming to invade your home and if you don’t stop them they’ll take over and your people will be oppressed for a thousand years’, and that victory feels better thinking it could have actually gone the other way, but nobody actually wants to walk away from that campaign being like, ‘Yeah, I guess we lost, I guess the elves won’. Maybe if you’re in a scenario where you can immediately be like, ‘Okay guys, now we’re running part two of this campaign, where now you’re going to fight the elves from the inside’, or something. We did that back in high school once, we had a fight, a big boss fight, that we’d lost, we all died, and then the GM immediately started a new campaign where we were going to take those guys down. But a), that’s a really big if that your players will be willing to stick with you for another campaign, and b), if you’re a grown-up, can you really assume that everyone is going to be available for a second campaign? Let me get all my memes out about how no one is ever available for a session.
So anyway, there are just a bunch of tricks that you can do. And one of them, a big one, is to fudge your dice, just fudge your dice.
Wes: Yep.
Chris: Lie!
Oren: Yeah, just lie, straight up lie if you have to. And this is what the GM screen is for, and most players will never question it. Most players, especially if they’re already hurt, if they’re already at 10 hit points out of a 100, they will very rarely be like, ‘Hey, did that troll really miss me?’ It’s very unlikely for that to happen.
Chris: You’re not going to look a gift horse in the mouth, right? The players want to succeed.
Oren: Yeah.
Wes: I think the other component with that is- and I’ve experimented with this with varying degrees of limitation, like success and failure, relating to attacks and other things, but yeah, the screen is a good thing to make use of, and I think making sure that you’re taking certain rolls into your hands even for that player, I don’t think that players should be rolling for everything related to their character, because if I ask Oren to make a perception check and he rolls and it’s low, he knows that he went low, and Oren’s a fantastic role player, but he still knows that he rolled really low. Right?
Oren: Yeah.
Wes: But if I made the roll for him and then we talk about what he saw, he has no reason to really question the believability of that. Like he could, but he doesn’t see the result. So I think if you can negotiate between what a character might reasonably be aware of knowing- like if you’re making a weapon attack and you roll low, that’s easy to know that your character would be aware of that. But for certain things I think it can keep that trust of the players not questioning things if you adopt it as more of a consistent habit.
Oren: Yeah, no, I’m a big fan of, if you are rolling the kind of spot check where a failure just means that if there’s something there you didn’t see it, I generally think that that should be rolled secretly, just because it’s just kind of a pain to try to role play: I don’t know I just failed a spot check. Like, it’s just, I don’t want to do it. I would rather just not.
Wes: Completely agree.
Oren: And there’s different ways of handling that, right? If you’re playing a fail forward system, the results of failing a spot check might be immediately apparent, at which point you wouldn’t really need to worry about it. But if this is like, ‘Did you spot the thief that’s following you?’ It’s like, ‘Yeah, roll secretly for that please. Make it easier for me to role play that I don’t know there’s a thief behind me.’
Another thing that you can do, early on- in standard fiction you would kill a major character early Ned Stark style. That doesn’t really work in role-playing games because we all know how sharp a difference there is between a player character and an NPC. But if you set up situations early where failure is actually a thing that can happen, like if the early conflicts are not about whether your people get enslaved for a thousand years, but are about over who gets a magic sword, and the bad guys actually win a few times, then this teaches your players that failure is possible. Failure can happen. And then when you get towards- and this also, incidentally, works in regular fiction, please take note, it’s called season three of Buffy the Vampire Slayer, where the bad guy is actually able to win a few times because the stakes of the conflict are not life or death for the main characters- and so then when you get to the end or some kind of climactic fight and you tell the players, ‘Yep, you guys sure pulled that off by the skin of your teeth, don’t look at my dice’, they will be like, ‘Yeah, we absolutely did, we got there, we are so great’, and they will feel fantastic and they will love you and you will be their best friend forever. Best friend offer not valid outside the United States.
Chris: I do think that it’s definitely easier to pull off kind of good stakes but not accidentally kill people in some systems better than others. It’s really hard in D&D because in a fight there’s really just nothing on the line other than hit points. And it’s hard to translate that hit point conflict into something else, right? Whereas if you have something that has other consequences or, I think that a lot of times, if you’ve got a story, an ongoing story for your campaign, what really the players are fighting for, at the end of the campaign, is a story outcome, not necessarily their lives, and with story outcomes it feels more realistic that they could fail, because the characters won’t die, but also it’s easier, I think, to have a fallback option- other consequences if players roll badly, right? Like maybe if a player rolls badly, the other player can make up for it, but now the whole team is kind of behind and has to work uphill to get the story outcome they wanted because there was failures at certain points, and etcetera, etcetera. Definitely easier if you have a system where you’ve got conditions, like in Mouse Guard in Torchbearer.
Wes: Yeah. And I mean, so D&D, from a design perspective, D&D style systems, you’re absolutely right, where the combat is about who can make the other guys’ hit points run out faster so they die, does have that issue where it’s just hard to make combat about anything else other than kill or be killed. At least with fifth edition D&D, a character is very unlikely to actually die. The mechanics have been made such so that you fall unconscious, at which point the rest of your party can win the fight for you. Now, of course, that doesn’t always help, right? If your party is just losing and there’s just no conceivable way for them to win the fight and then you have to try to figure out how do I tell them that they should retreat or something, right? And that’s a little bit trickier- this is one of the reasons why I don’t play D&D very often- but you can do that. There are a couple of things you can still do from a GM’s perspective there. One that I really like is to have some allies or weird, mysterious magic items or whatever that you’ve set up ahead of time that no one’s really sure what they can do. And it’s like, ‘Yeah, you’ve got a weird staff that talks to you sometimes, and it seems to be magic, but what kind of magic really?’ And then if it really seems like the character is about to die, then you can reveal that, ‘Oh, actually this staff has life steal powers, and it keeps you alive’, or, ‘Oh, actually, those weird monks that gave you food earlier are here to help you, and they know kung fu’. And this obviously has its limits. If you go too far with it, then the PCs will be like, ‘Oh, you mean every time I die, I get a new power, awesome!’
Wes: But it’s a good option to consider, and I think the other thing that we could advise GMs is don’t forget that your monsters or your villains, they generally are thinking, rational- well, degrees- of, you know … There’s no reason why you have to make the antagonists kill the players. If you have a very powerful villain, maybe it’s a test, right? If it’s going that way, they could just pull back and then suddenly that goblin leader decides to not kill them and leave to report on their strength. And suddenly you’ve turned what was just the most powerful goblin on the board into a character in their campaign that they could know about or find out about later. Right? It’s okay to not just always pursue kill counts with your antagonists.
Chris: Yeah. It goes back to giving your bad guys goals that don’t require the protagonists to die, which is just a very useful thing in storytelling in general. I do remember, so, Oren, one time you ran an L5R campaign, and L5R has exploding tens, and those create some very random results that make it really hard to not have random death happen in the fight.
Oren: This is true.
Chris: So we had this L5R campaign and we had a couple of times when a player character just rolled into the red, died right off, and you just, basically by fiat, decided they weren’t dead but then imposed a story consequence that made it feel satisfying. So basically we became haunted.
Oren: Yeah, now you got ghosts.
Chris: Yeah, now we had some spirit head who had intervened and saved us, but now we were this kind of malicious spirit hanging around and talking to us all the time, and possibly wanting to take over our bodies, that was the thing that I had to deal with.
Oren: Yeah that was super great fun.
Chris: Yeah. But no, it was a really good thing. I think that if you have to intervene and keep somebody from dying, it definitely is more satisfying if it doesn’t feel like the death was just hand-waved away, if it feels like there was some consequence for that happening, in this case a really interesting story consequence, which is honestly more like a reward, but even so in the story it feels like you have more problems as a result of this happening. And this just inherently makes it feel like it wasn’t erased, like it still matters.
Oren: Right. And L5R is definitely one of those systems where it’s more challenging to not kill player characters because they’re so fragile, and just damage rules can get really ridiculous in that game. So I was definitely tap dancing a little harder than normal when I was running that game. Because I was like, okay, I really need these characters not to die because I have such cool stories planned for them. Another solution that I like the most is to use systems where random character death just doesn’t happen.
Wes: Yeah.
Oren: This gets a little tricky because, again, players want to win by the skin of their teeth. They want to feel like they could’ve lost. So you kind of move the goalposts a little bit, once random character death is taken off the table, but trying to figure out how to create the illusion that they could have failed is just a lot less high risk than the possibility of characters dying. Because a character dying is, in most cases, basically a non-starter. If you’re trying to make a campaign, just randomly killing a character will almost always be bad. If you’re going to do that, it’s something you really need to think about ahead of time and have something the player is actually interested in at least as a possibility.
So I’m actually a big fan of Tenra Bansho Zero, the way that it does it, where your character- in Tenra Bansho Zero you get bonus dice when you’re wounded because you’re super cool and have blood running down your face and whatever, and the last box on your character sheet is your death box, and if you check that off, you get even more dice, but now you can actually die, and that’s at least giving the player agency in it, that’s having the player be like, ‘Yeah, I’m willing to die in exchange for more dice.’ Now that’s not a perfect solution, because people make bad decisions, and sometimes the player might do that thinking they’re not going to die, but then die and not want to. So it’s not a perfect solution, but I like it better than the D&D model. I’m also just a big fan of more abstract combat systems that have objective-based fighting, where character death, if it’s gonna happen, can come down to some really dramatic plot result, as opposed to too many critical hits by a goblin.
Wes: Right.
Chris: Yeah. Certainly the issue here is not really that characters die, it’s that they are not planned in any way.
Wes: Yeah.
Chris: Because if you plan, then you can get the player on board. And some players, in some situations, are perfectly fine with their character dying in some glorious fashion, and you can end their arc, but if it’s random, if it can happen when you don’t plan it ahead, then it’s impossible to control or make it fun in any way.
Wes: And with a little planning it really doesn’t take too long- with having your players flesh out their bonds and their connections and their place in the world, you suddenly can have more story-based threads for things that they would care about rather than a worry that their character might die.
Chris: Yeah.
Wes: Yeah. Oh, and here’s another one, we’re almost out of time, but here’s another, just a fun- a useful on the ground trick, if you’re running a D&D style combat system, is to just have a running total of your player’s hit points available to you, because if you know that some player is about to die, it’s usually pretty easy to come up with an in-character reason for why they aren’t getting attacked. Because even though mechanically speaking you still fight just as good at one hit point as you do at full hit points, you can say that the troll looks at you and sees that you’re almost dead and you’re not a threat and so goes to attack someone else. And that’s a tactically unsound decision within the confines of the game but narratively it makes sense, right? And players are usually willing to accept that. Again, some aren’t, some will really press you on this because not all players are good people, but if that happens, that’s a bit of a-
Chris: Just kill their character.
Oren: Yeah, I guess you could kill them.
Wes: Dead.
Oren: Yeah, if they’re really pushing you, you can just kill their character, then maybe that’ll make everyone else think character death is a possibility. All right, well we are out of time, so we are going to have to go until next week. But before we do that I want to thank two of our patrons, Kathy Ferguson, who is a professor of political theory in Star Trek, and Ayman Jaber, who writes about fantasy stories at thefantasywarrior.com. For the rest of you, if anything we said piqued your interest, you can leave a comment on the website at mythcreants.com. Otherwise we will talk to you next week.
Chris: This has been the Mythcreants podcast, opening, closing theme, ‘The Princess who Saved Herself’ by Jonathan Coulton.
I love your show, but I take UMBRAGE at the idea that dice should be fudged to avoid character death.
If you find yourself having to ignore the rules of a game, then probably one of these is happening: a) either you’re playing the game wrong, or b) you’re playing the wrong game, or c) the game Is terribly designed.
If you find the game you’re playing has too high lethality foto your taste, maybe it’s meant to ve that way. Maybe it’s meant to encourage you to avoid combat at all costs like in Cthulhu Dark, where engaging a creature in combat always means you automatically die. Maybe the survival element is part of what makes it fun, like with old school D&D (after all, if character creation Is little more than rolling 3d6 six times o
in order and picking one of seven clases, you should be back playing in no time).
If character death bugs you so much, why not play games in which character death is not an option? Example: Icons, a superhero game based on FATE, specifically states that in the superhero genre death isn’t ever permanent, so your character will be back in a couple sessions. It makes a lot more sense to just play a nonlethal game than to roll dice and then just do whatever you like (which is pretty much just playing pretend but with a plastic rattling pantomime).
Okay, that’s out of my system. Love the show, I just really strongly disagree on that one point.
DO NOT fudge your dice rolls. If you fudge rolls to save characters (or to kill them) you are just a lying cheater.
There are an infinite number of ways to have the game you’re playing create the kinds of results that you want, if you set up the initial conditions properly. If your group decides that character death is off the table (this is a conversation for session zero), then you should arrange the game beforehand to prevent this from happening.
I personally don’t fudge dice, but my players are aware of that fact and they like the unpredictability of taking any role as I comes.
Our safety feature is that character death (or severe injuries) can be taken if the player likes but are not a must – except the deliberately overextend.
And yes i had fights where the opponents lost through bad look but this adds to my players enjoyment as they see that they are not the only ones with bad luck.
I think the vail between in-time and out-time are quite blurry in my groups and I have nothing against fudging dice, its a matter of taste i guess.
My personal take on this is that as a GM you are not a slave to the rules and mechanics of the system you are running but you should try to follow the rules as much as possible. Personally I am ok with fudging some rolls in rare instances or possibly to just not roll if you’re try to drive a very cinematic scene for example think of the chase scenes in Lord of the Rings or the Hobbit when they are running from the goblins/balrog. If you are trying to create a scene with that sort of tension I think it’s perfectly reasonable to simply allow the entire success/failure state rest on the skill checks of the players and as the GM not roll any dice at all. I will run these as extended skill checks where the players basically need to not get say 3 failures in 5 rolls assuming they don’t fail the skill challenge then when a horde of goblis try to attack them when they run by I don’t roll an attack for the goblins they just fail their attack because the player hasn’t failed 3 of their tests yet. Basically bust because something might feel like a combat scenario doesn’t mean you need to use the mechanics for combat in that situation this keeps the tension high without allowing the mechanics to dictate how you are telling the story.
I will say you should avoid changing dice rolls for things that are directly detrimental to your PCs whether it is in their favor or not. That said you again you are the GM you don’t need to be a slave to the mechanics of the system. If the rules say the PC should be dead it is completely within your power to say “fuck that you wake up in a dark room tied to a chair with a orc standing over you and now need to figure out how to escape.” Allowing your heroes to be flat out killed because they lost a fight when you don’t want them to die is a result of you as a GM lacking the creativity to apply appropriate non-lethal consequences to their failure. It’s the opposite side of the coin to something that plagues every GM which is when players think that the only solution to any given problem is stabbing it in the face and being murder hobos. Your NPCs don’t have to be murder hobos any more than your PCs do. A big part of what I do as part of my prep as a GM is creating fail state encounters. If I have a 12 foot wide chasm that I know they will need to cross as part of an encounter I have a “what if” encounter ready for them should one or more of them fail to make it across and end up falling into said chasm if none of them fall then that what if encounter gets tucked away so I have something that’s already prepared for the next time I have a chasm or “bottomless pit” type hazard. If they are fighting a tough group of bandit I have a what if they lose to the bandits encounter ready where maybe they are captured caged and sold as slaves. This derails the adventure and gives the players a story based consequences to their actions that then now need to resolve.
In cases of a partial failure for example two out of four party members hit 0 hp in the encounter but the other two finish the fight those two PCs that “died” may just be gravely wounded and the surviving two members now need to desperately attempt to keep them alive by finding medical herbs while the two that fail find themselves in an astral form where theier spirits must fight to keep a connection to their physical bodies long enough for their friends to find those healing herbs and get them back on their feet. Maybe they have a near death experience and get sent to hell or some other outter plane only for their patron deity to send them back to complete some holy mission. There was even a time when one of my PCs should have died and instead I told him he got his sword arm mangled which got infected and they to have the arm amputated. So they had to spend a fair amount of time learning to fight with their off hand and eventually got a magical construct to replace their lost arm and ended up being ambidextrous after that and had a sweet cyborg arm. So this is one of my favorite things to do when my players fail combat as there are some serious consequences for failure but it eventually gets to the point where had they not had that failure they wouldn’t have got some really cool thing later down the road so my players learn to embrace failure and understand that it’s not always a bad thing and can lead to some pretty awesome upsides later.