The Delta Theory of Meaningfulness

[AceOfFools]

What's important is that failure/death be, at some level, the result of decisions made, and preferably the result of multiple decisions made.

…

1. Decisions should have consequences
2. Consequences (especially negative) should be the result of decisions.

This is exactly my point.

In the first attack a group picks a direction based on a die roll, fails some whatever-they-called-spot-checks-in-the-80s, two more bad rolls later, and characters was dead.

In the second, the party did decide to head into some cool sounding ruins, but the ambush was considered otherwise unfair. One “we have to pick something, I guess” and one bad roll, and dead.

The beholder I fought was the Boss Monster of the scenario, planned from the campaign start. When we started a game an out hunting monsters to save the world (or, y’know, one small city), we were deciding to take on this monster. We managed to find out there was a beholder, pack on every advantage we could—including an ambush, and then despite every decision we did made—including forgoing a sneak attack due to fighting defensively—and two unlucky rolls and I was dead anyway.

The fact that a few die rolls can make a character die, no choices beyond “let’s play the game” can cause dire consequences.

Playing under rules where “your character cannot die, unless you agree it’s the narratively appropriate outcome,” makes all character deaths absolutely meaningful choices of the player by default.

And as long as you’re playing under the paradigm of wanting for the opportunity for high Delta—to make meaningful choices, rather than wanting to “win”, you can make that most meaningful choice.

My favorite character death was the result of a failed strength check in a DnD 4e game. I asked the GM if, instead of the child being pulled into the hell portal, if I could push between the child and the demon instead drag my character to hell instead. He said “…yes”, and we later decided that the choice for the character would be more meaningful if, instead of the planned hell rescue followup quest, we left that character to his fate.

This is was an extremely high Delta moment, one that only happened because of a single unlucky roll (mostly). It could only happen because the system didn’t force it (the normal result of a failed check was the child was pried from your hands).

[kyoryu]

This is exactly my point.

In the first attack a group picks a direction based on a die roll, fails some whatever-they-called-spot-checks-in-the-80s, two more bad rolls later, and characters was dead.

In the second, the party did decide to head into some cool sounding ruins, but the ambush was considered otherwise unfair. One “we have to pick something, I guess” and one bad roll, and dead.

The beholder I fought was the Boss Monster of the scenario, planned from the campaign start. When we started a game an out hunting monsters to save the world (or, y’know, one small city), we were deciding to take on this monster. We managed to find out there was a beholder, pack on every advantage we could—including an ambush, and then despite every decision we did made—including forgoing a sneak attack due to fighting defensively—and two unlucky rolls and I was dead anyway.

The fact that a few die rolls can make a character die, no choices beyond “let’s play the game” can cause dire consequences.

Playing under rules where “your character cannot die, unless you agree it’s the narratively appropriate outcome,” makes all character deaths absolutely meaningful choices of the player by default.

And as long as you’re playing under the paradigm of wanting for the opportunity for high Delta—to make meaningful choices, rather than wanting to “win”, you can make that most meaningful choice.

My favorite character death was the result of a failed strength check in a DnD 4e game. I asked the GM if, instead of the child being pulled into the hell portal, if I could push between the child and the demon instead drag my character to hell instead. He said “…yes”, and we later decided that the choice for the character would be more meaningful if, instead of the planned hell rescue followup quest, we left that character to his fate.

This is was an extremely high Delta moment, one that only happened because of a single unlucky roll (mostly). It could only happen because the system didn’t force it (the normal result of a failed check was the child was pried from your hands).

I think there's also something in there about consequences either being extremely obvious, or the result of multiple decisions.

Like, in your case, you basically knew the choice you were making, and made it. That can be a one-and-done.

In the second, just a couple of bad rolls (at best), for things that should have been moderately benign, led to death. (Since you're adventurers, and you're supposed to adventure, choosing to do baseline adventurer stuff shouldn't really be considered an obvious indicator of death, otherwise boring game). That's not as cool. Failing a spot check can put you in a worse position, but shouldn't just straight up murder you.

[Easy e]

Some of my favorite games are explicit that your character can not die. Note, that doesn't mean they succeed either. There are many fail states, often dying is the easiest one and doesn't create tension in and of itself. This fixation on dying as the only fail state is a clear D&D-ism.

Dying only matters if your death leads to some larger stakes or outcomes. If you dying just means roll-up a replacement, that is not inherently tension, it is just annoying. I really hate dying for no reason.

The time I died when I got mobbed by Goblins and stabbed to death in session 1? Just annoying that I bothered to put anytime in making a character beyond the numbers in the first place. The time I made a character and got killed on my first die roll of a d6 Star Wars game, just annoying. Dying in a long running campaign thanks to a "Save or Die" in the middle of the campaign, just annoying as I was the only one going through "new player" mode in a well-established group. In those situations, I learned to not care about my character as they were just an expendable game token, no more, no less. Who cares if I walk face first into a death trap? No one. If I LeRoy Jenkins a dragon by myself? No one. Just make a new token! Boring.

As a player, I have no problem dying as a character for some stakes or reason though. In a game where I can not explicitly die, I actually had my character die to prove a larger spiritual, moral, and ethical point. In a different game, my player purposely performed a "last stand" so others could escape and save the world later when they would not have been able to otherwise. I had no issue with those, and it really drove the story and character forward for everyone else.

Death has no inherent stakes or is a better fail state in a game.

[KorvinStarmast]

I think there's also something in there about consequences either being extremely obvious, or the result of multiple decisions.

That cancels out mystery or hidden antagonist games, like the Rashasa crime lord ...

This fixation on dying as the only fail state is a clear D&D-ism.

Incorrect. I'll offer two examples that are not D&D.
1. Traveler, the original. You did a lot of stuff and you took a lot of care not to die.
2. Mothership (recent sci fi horror). Death is gonna happen. We have had two PCs die already.

[Easy e]

That cancels out mystery or hidden antagonist games, like the Rashasa crime lord ...

Incorrect. I'll offer two examples that are not D&D.
1. Traveler, the original. You did a lot of stuff and you took a lot of care not to die.
2. Mothership (recent sci fi horror). Death is gonna happen. We have had two PCs die already.

More games than D&D can have a D&D-ism. Horror, especially survival horror; is a different beast completely.

However, that is completely besides the point of what I was discussing. Death is not an effective fail-state as there are very few stakes in dying in an RPG. You just make a new character and "re-spawn". Nothing is really lost, it is just a distraction, annoying or a hassle.

[Atranen]

Making a new character, having to integrate them into the party, possibly losing exp and levels, losing a connection with someone you've developed...ime there are substantial stakes to character death in RPGs.

[Vahnavoi]

More games than D&D can have a D&D-ism. Horror, especially survival horror; is a different beast completely.

Character elimination as a loss condition is ubiquituous in all sorts of games, many predating D&D, so calling it a D&D-ism is a stretch. There's a very simple reason for that: in a game where a player primaly participates through their character, character elimination is equal to player elimination, and a natural stopping point for that player. Note that "character" and "token" are exchangeable in the previous sentence. Equating loss of character with loss of a token doesn't somehow prove character loss isn't a consequence - losing game tokens is a basic game consequence.

However, that is completely besides the point of what I was discussing. Death is not an effective fail-state as there are very few stakes in dying in an RPG. You just make a new character and "re-spawn". Nothing is really lost, it is just a distraction, annoying or a hassle.

This is equivalent to saying losing game tokens isn't an effective fail state because you can just start over if you lose. The flaw in logic should be obvious.

The actual way games work is that the annoyance of losing your tokens and having your participation end - losing a game, in short - is the consequence you're trying to avoid. Don't like spending time making a new character? Play better so your characters last longer.

The real flaw you and AceOfFools observed has more to do with a game ending randomly - it isn't possible to play better because the game wasn't lost by any action on the player's part.

It's worth noting that a lot of things that seem random or unable to be counterplayed in a single game, might not be so across an extended game where the situation is repeated multiple times. A basilisk is a great example. Yes, the first time you encounter a basilisk, it might be an out-of-nowhere instant game over. The second time, you know a basilisk might be in the game, and can employ various counter-strategies (divination spells to find out where it is, amulets to protect from petrification or magical darkness or indirect fire to stay out of line-of-sight, etc.).

It being easy to replace a character goes hand in hand with extended play, where such learning through trial and error is expected and normal.

[warty goblin]

If you want rolling the dice to be meaningful, i.e. have a potential impact on the outcome, then you have to be able to win or lose by luck. Saying you shouldn't ever lose because of bad luck is equivalent to saying you should never win by luck and also that you shouldn't have luck at all, at which point you're better off playing chess. Basically if you include meaningful dice, you're making the meaningful choice to win or lose randomly.

[Vahnavoi]

@warty goblin: it isn't quite that simple. A game can have meaningful randomness while still having winning and losing be deterministic based on player ability.

A simple example would be a randomized hexcrawl, where each possible hex has a distinct puzzle - but each puzzle is individually solvable and deterministic. Every played game can hence have different sequence and stopping point (dependent on player skill) without "luck" being real determinant.

This said, most popular dice-based engines aren't set up in this way. They are instead straightforward gambles on game resources, character life included. In such games, there is a hard statistical cap to frequency of success and even the best possible game strategy will lose some amount of time. For example, every published version of D&D is in this boat.

The spirit of your statement is still correct - don't play a game of chance if you can't accept possibility of losing by chance.

[KorvinStarmast]

Making a new character, having to integrate them into the party, possibly losing exp and levels, losing a connection with someone you've developed...ime there are substantial stakes to character death in RPGs.

Yes.

Character elimination as a loss condition is ubiquituous in all sorts of games, many predating D&D, so calling it a D&D-ism is a stretch.

As but two examples: Diplomacy(no dice) and Risk(yes dice).

The actual way games work is that the annoyance of losing your tokens and having your participation end - losing a game, in short - is the consequence you're trying to avoid. Don't like spending time making a new character? Play better so your characters last longer.

I now await the Greek Chorus complaining that player skill shouldn't matter in an RPG ...

It's worth noting that a lot of things that seem random or unable to be counterplayed in a single game, might not be so across an extended game where the situation is repeated multiple times. A basilisk is a great example. Yes, the first time you encounter a basilisk, it might be an out-of-nowhere instant game over. The second time, you know a basilisk might be in the game, and can employ various counter-strategies (divination spells to find out where it is, amulets to protect from petrification or magical darkness or indirect fire to stay out of line-of-sight, etc.).

It being easy to replace a character goes hand in hand with extended play, where such learning through trial and error is expected and normal.

True with D&D, and also true with games like Diablo (the original, a dungeon crawl on a computer) that you can play through more than once with more than one starting character. The player likely will remember a few of the nastier surprises (the spitting dogs in the catacombs for example) when they try the game with the next choice. (Say they did their first play through with a warrior, their rogue attempt will have some player experience to aid them in approaching certain zones of the game).

A game can have meaningful randomness while still having winning and losing be deterministic based on player ability.

Casino craps being one such.

The spirit of your statement is still correct - don't play a game of chance if you can't accept possibility of losing by chance.

Craps again.

[Easy e]

You know what. I won't convince anyone else of my position that death doesn't matter in RPGs, and that is fine. The way I think about games and many people on this forum think about them is different. That is why I learn a lot about RPGs by reading these different takes and thoughts. Thank you all for sharing.

I will say that from my perspective the only "Meaningful" decision in an RPG game is when you decide to play and when you decide not to play. The rest doesn't matter. By playing you are embarking on a collaborative experience, and there is no winning or losing, there is only experiencing. The playing was either worth your time or it was not, and you often do not know which until it is all over.

The Play is the thing.

[icefractal]

True with D&D, and also true with games like Diablo (the original, a dungeon crawl on a computer) that you can play through more than once with more than one starting character. The player likely will remember a few of the nastier surprises (the spitting dogs in the catacombs for example) when they try the game with the next choice. (Say they did their first play through with a warrior, their rogue attempt will have some player experience to aid them in approaching certain zones of the game).

Maybe this is the disconnect for me? I've heard about these megadungeons where the party (or even different parties over time) kept returning to the same one, building up knowledge and slowly going deeper. But I've never seen that in practice. Even the more dungeon-crawl-heavy games I've been in, we went through a given dungeon once, and that was it.

So let's say I walk around the corner, foolishly thinking I can walk around corners with my eyes open, and get stoned. The rest of the party either avoids or slays the basilisk, and then we never return there. Unless it's a particularly large dungeon, my new character probably shows up after we're done with the whole thing.

Now true, basilisks in particular could leave telltale stone remains around (although so does a trap that petrifies people, which is something you'd definitely want your eyes open to avoid!), but there's also things like Bodaks which just leaves corpses. You're saying that if you saw an old corpse on the ground, your instinct would be "better close my eyes and navigate by touch/sound"?

[Vahnavoi]

Nope, that's not what's being said. Megadungeons are also tangential to the point - even if you only go through each dungeon once, that doesn't mean you will only meet a given enemy once.

So the first time you and your friends meet a basilisk, one of you suffers an instant game over - the rest deal with it or flee from it. Okay. Then you get to the next dungeon. Will you now approach it the exact same way as before, forgetting everything that happened with the basilisk?

For the record, it's possible to overpopulate a game with weird monsters, so that it becomes impossible for players to really learn anything from past experience. Some game masters do this on purpose because they have counter-productive ideas about player knowledge. If that's your experience, you have my sympathies, but please stop moving the goalposts by going "oh, what about that other monster then?".

[icefractal]

Then you get to the next dungeon. Will you now approach it the exact same way as before, forgetting everything that happened with the basilisk?

Please describe to me the approach you take to dungeon delving, knowing that basilisks exist.

For the record, it's possible to overpopulate a game with weird monsters, so that it becomes impossible for players to really learn anything from past experience. Some game masters do this on purpose because they have counter-productive ideas about player knowledge. If that's your experience, you have my sympathies, but please stop moving the goalposts by going "oh, what about that other monster then?".

I don't really think "two different monsters" is overpopulation, but my point was just that if someone was taking the stance that "all basilisks should be signposted by the petrified bodies of former victims" (what about wandering ones though?), then there are other insta-death monsters which don't leave a distinctive calling card like that.

Incidentally, Bodak is another creature that got changed to "not insta-death" in PF1, and I'd say the change is an improvement.


I wouldn't say that insta-death monsters reduce delta per-se (after all, it's not like you couldn't already lose and die entirely via bad luck in most systems, insta-kill stuff just makes it more likely), I just don't think they're usually a benefit for the game.

[warty goblin]

If one's philosophy is sufficiently simulationist/non-PCs-are-special, I don't think there's any inherent problem in some situations not having a solution and those situations not being the result of any obvious or predictable mistake.

[Devils_Advocate]

I agree that it's not inherently bad for some goals to prove unexpectedly unachievable, even by design. Seeing how characters cope with something like that can be interesting. I guess that the main question there is how often feels like "too much": At what point does it seem like our heroes should really just pick a new career?

Oh, and beware railroading, of course. Just like a GM shouldn't commit to a problem having exactly one solution, a GM shouldn't commit to a problem having exactly zero solutions. If the players find a way to do the "impossible", reward their ingenuity.

If you want rolling the dice to be meaningful, i.e. have a potential impact on the outcome, then you have to be able to win or lose by luck.

Potential results can significantly differ from each other in ways that don't make one clearly and straightforwardly "better" or "worse" than another, and that's often more interesting than degrees of clear-cut "success" or "failure".

Saying you shouldn't ever lose because of bad luck is equivalent to saying you should never win by luck and also that you shouldn't have luck at all, at which point you're better off playing chess.

Or an RPG without random elements, which feels like the more relevant alternative in this context.

On the highest level, it’s a game. No matter what happens, when we’re done playing, we go on with our lives. The choices had no effect outside the game. Like watching a movie, it’s a pleasant way to turn 7:00 into 10:00.

Stories can have lasting impacts on people. I, uh... don't think that it's really normal to engage with works of fiction, or with films and RPGs specifically, in an exclusively transient fashion? I have no statistics on this, but my suspicion is that you're the weird one there. Or perhaps I've misunderstood you.

We care if we win a game of chess or Monopoly.

I don't, and I'm sure I'm not alone. And I doubt that winning is of primary importance to a majority of gamers.

Trying to win is the most important part of the game.

Winning is kinda secondary.

I don't know if I make sense. On the same vein, if you meet Buddah on the road, kill him for 5,000 xp.

In the original game of D&D, in which characters can and will die if they lose a fight, all choices were meaningful, and obviously so. If our characters don’t defeat (or escape) these ogres, they will die. And nobody ever had to ask if our choices were meaningful.

But choices are only meaningful insofar as the variance in consequences is meaningful, not just to the extent that that variance exists. And the more regularly that player characters die, the more that players are discouraged from putting any actual, y'know, characterization into them (because it's not worth the effort to write up a detailed backstory nor personality for a character who's only going to last a few hours), instead of churning out a stream of nameless, faceless store brand adventurers whose only notable motive is the acquisition of wealth (as the default motivation for going into monster-infested dungeons is "that's where the treasure is").

Having to play Generic Fighter #378947 instead of Generic Fighter #378946 hardly seems like a terribly meaningful consequence.

But without a weird result from the dice, we *couldn't* just play an identical character. When my first Fighting Man died, I rolled up a character with 18 CHA. This one was a paladin (the first one our group had ever seen).

The cliché of “Rolling an identical character to the one who just died is" couldn't happen in the first several years of my playing – not when we rolled 3d6 in order.

"Having to play Generic Fighter #378947 instead of Generic Fighter #378946 hardly seems like a terribly meaningful consequence."
"But they're not the same. One of two Fighters might have higher bonuses to his attack and damage rolls, while the other has higher hit points."
"So what, though? Why would I particularly care?"
"Um, because those numbers determine the outcomes of combats? They decide whether characters live or die!"
"Yeah, but dying just means having to play Generic Fighter #378947 instead of Generic Fighter #378946, and that hardly seems like a terribly meaningful consequence."

In the same way, we could call whether our PC dies a "cosmetic choice". If he dies, I’ll just create another one. What difference did it make?

If the "character" was Generic Fighter #378946, no difference of note. But if the character was an ousted prince trying to avenge his murdered father, well, now we're talking. Then having him replaced by a previously-unmentioned brother upon his death would likely feel dissatisfying and jarringly silly, cheapening both the character and the story. Heck, even if such a brother does show up, there will probably be at least some focus on the differences between the two.

It's also potentially dissatisfying for that character to die to a random bandit ambush, so the "danger" is that GM will consequently avoid that. But that's not actually a problem, in the same sort of way in which it's not a problem if a character can't be killed in one blow by a particular adversary. One might argue "Having a lethal weapon swung at you should always carry the chance of death!" And the counterargument is that we're just not interested in anything like that sort of realism. Players don't just prefer that their characters live rather than die. Many would rather have them die to a goblin attack than to gangrene. And, similarly, to a wyvern attack than to a goblin attack, and so on and so forth. Cutting out the crappiest crap doesn't mean that some consequences aren't worse than others, nor does it drain events of meaning.

(And to anyone asking "If goblins aren't 'worthy' of killing the main characters, should they even be fighting such pathetic enemies?", I congratulate you on your keen insight.)

Taking character death off of the table entirely is obviously pretty restrictive, eliminating many possible sources of tension. But not having it take up half of the table makes more room for other stuff. One bit of information, to borrow from NichG, can only say so much.

So the real answer is to care about the cosmetic things within the game. Going up a level, gaining a magic item, using illusions instead of summoning, etc. Those things matter to my character. If I’m role-playing him, then they should matter to me.

There's a certain tension between "I don't want my character to die" and "I want my character's life to be significantly threatened". Or more broadly, "I want my characters to achieve their goals" and "I want my characters to face obstacles that reduce their likelihood of success." There are lots of different potential ways to address that tension, with various tradeoffs.

[Vahnavoi]

Please describe to me the approach you take to dungeon delving, knowing that basilisks exist.

I already did. Again: use information gathering methods, such as divination spells, to check if there's a basilisk in the next destination, buy an amulet that protects from petrification, or use indirect fire or cover of darkness to fight it without entering its line-of-sight.

Notice the first step generalizes to all troublesome creatures. If I'm playing any old school dungeon crawling game, one of the priorities is getting some equivalent of Scroll of Detect Monsters & be watchful for any known troublesome creatures. In absence of such options, the most general strategy is to pay attention to specific context of the encounter (location, time of day, who was with it, etc.) to make guesses about where future troublesome creatures might be.

I don't really think "two different monsters" is overpopulation, but my point was just that if someone was taking the stance that "all basilisks should be signposted by the petrified bodies of former victims" (what about wandering ones though?), then there are other insta-death monsters which don't leave a distinctive calling card like that.

It's two different monsters so far, but you (or someone else) could repeat the same "what about this other monster?" until we're all sick of it. I'm telling to stop now because it doesn't matter. It doesn't even matter for the argument you're supposedly countering: if someone's of the opinion "instant death monsters should be signposted", then their answer to any "what about this monster that isn't signposted?" would presumably be "then add the signpost or don't use the monster". Nothing forces anyone to use any monster in a completely unpredictable manner.

I wouldn't say that insta-death monsters reduce delta per-se (after all, it's not like you couldn't already lose and die entirely via bad luck in most systems, insta-kill stuff just makes it more likely), I just don't think they're usually a benefit for the game.

Meanwhile, the point I'm making is that the benefit (or detriment) of any such monster depends on format of a game. They have different pay-offs as one-off elements than they have as repeat elements.

---

If one's philosophy is sufficiently simulationist/non-PCs-are-special, I don't think there's any inherent problem in some situations not having a solution and those situations not being the result of any obvious or predictable mistake.

While desire for simulation and ideas about character status might contribute, I don't think those are at the root of the matter. There's a far more basic game design question that could apply to a fully abstract game, such as a card solitaire. Namely:

Do I want to ensure that every permutation of my game is solvable?

It's worth noting that increasing hidden information and amount of randomness in a game both make it more difficult to ascertain that. For two popular card game examples, both Klondyke and Demon solitaire are notorious for having a lot of unwinnable hands and, more relevantly, it being extremely difficult to tell whether your hand is winnable until well into a game.

So, there's a point where the earlier question becomes equal to:

Do I want a game of imperfect information with genuine surprises in it?

and

"Can I deal with a game where, even if I play the best possible way, it still can't guarantee a victory?"

[Slipjig]

I don't agree that character death is what inherently makes things meaningful.

I mean, if we're talking AD&D, I was in an AD&D campaign where one of the players lost their character literally before they had a chance to do anything - rolled it up, joined our group, we had a random wilderness encounter, an enemy that won initiative one-shotted them.

AD&D was also a very different game, with very different expectations. But even in AD&D, that was some not-great DM-ing. Unless the players made some truly dumb decisions leading up to that, e.g. "The Magic-User will take point."

Death doesn't necessarily make things meaningful, but the absence of real risk cheapens the game overall. If everybody at the table knows that the DM has a "characters don't die unless the player agrees" rule or, even worse, the character CANNOT die unless they've already gone down X number of times (a la the 40k games or Candela Obscura), it's impossible to have any real tension in most combats. Sure, you can still have other stakes, but if the character isn't at risk, there's no opportunity for genuine heroism.

But you CAN create the tension even without risk of PC death, just by letting the players know that they CAN fail. And that you have a plan for what happens next in the campaign if they DO fail (and that they will enjoy it considerably less than the success path).

[Vahnavoi]

AD&D was also a very different game, with very different expectations. But even in AD&D, that was some not-great DM-ing. Unless the players made some truly dumb decisions leading up to that, e.g. "The Magic-User will take point."

What NichG described doesn't prove any pathological wrongdoing on their dungeon master's part - simply due to how die rolls are set up, literally every low-level character has some chance of dying before getting to do anything. This is likely true of every edition of D&D when played by the book, with more recent ones, it's just so unlikely that you'll never see it in practice.

[NichG]

AD&D was also a very different game, with very different expectations. But even in AD&D, that was some not-great DM-ing. Unless the players made some truly dumb decisions leading up to that, e.g. "The Magic-User will take point."

Death doesn't necessarily make things meaningful, but the absence of real risk cheapens the game overall. If everybody at the table knows that the DM has a "characters don't die unless the player agrees" rule or, even worse, the character CANNOT die unless they've already gone down X number of times (a la the 40k games or Candela Obscura), it's impossible to have any real tension in most combats. Sure, you can still have other stakes, but if the character isn't at risk, there's no opportunity for genuine heroism.

But you CAN create the tension even without risk of PC death, just by letting the players know that they CAN fail. And that you have a plan for what happens next in the campaign if they DO fail (and that they will enjoy it considerably less than the success path).

What I'm pushing back against here is the tendency to focus on 'delta' coming from things like combats, failures, etc. There's nothing wrong with having character death be a thing or failures be a thing, but its like spiciness (heat) in cooking - if your idea of how to make meals tasty was too strongly focused on 'what is the correct spice level to use?' you're really limiting yourself.

Without any failure, risk, death, combat, etc, you can still have lots of meaningful choices. The richest, most meaningful choices actually tend to occur outside of those sorts of conflicts, because 'what you want' is pretty clear once initiative has been rolled - mostly you want to kill the enemies, not die, and use as few permanent resources as possible. The consequence for screwing up is strong (e.g. it has an inherently very high saliency, like spicy food getting your attention), but outside of that its more one-note than other things.

Compare that with a choice like 'what do we do with this old castle we cleared out near the border town?'. Turn it into a base for the PCs? Okay, how would you like your base, what sorts of features do you want it to have, etc? Not just aesthetic - you'll be using those resources for the rest of the campaign perhaps. Turn it over to the local nobility in exchange for favors? Okay, interesting, which noble house in particular do you want to make stronger and more influential, and what kind of favors are you looking for? Just leave it sitting there abandoned? Okay, lets figure out what sorts of things might move back into it over the next year or two, that will probably have some consequences for the region. Break it down for raw materials and sell them? Okay, that takes some time and becomes a local endeavor, so this is what we're doing for the next few sessions - interactions with the local Masons' guild, Merchants' guild, finding buyers, finding ways to turn the new influx of wealth into stuff you actually want (and deciding what you actually want!), etc.

Those choices don't have the sting of losing progress when you make them 'wrong', but they can be very complex and have ongoing effects and can make lots of different things about the world matter. If combat and death is the spice in this metaphor, those sorts of 'what do we want?' choices are the umami, the aromatics, etc.

[Vahnavoi]

We could ditch the food analogy and more explicitly talk about different aspects of gameplay. It's possible to break down what makes a game "fun" to a player into several widely recognized aesthetic categories:

1) Sensory pleasure, which can be further broken down by sense: sight, hearing, touch, taste, smell.

2) Self-expression: the ability of a player to bring up their own ideas and identity through medium of the game.

3) Discovery: learning and finding out new things.

4) Challenge: competition and testing of skill.

5) Narrative: building and experiencing a resonant story.

6) Social: enjoying company of and building up relationship with other players.

7) Submission: losing one's self to repetitive routine.

etc.

So there are at least this many dimensions of possible change when a player has their move.

True cosmetic changes are those that primarily affect sensory pleasure with a side of self-expression. A simple example would be having a choice between a couple of novelty chess sets or maybe painting the chess pieces yourself. The change does not move the game's state forward in any way.

For contrast, character elimination and consequent player elimination are traits of challenge-focused gameplay: the game state changing or threatening to change is a signal for the player to change what they're doing. As already noted, this ceases to work when the elimination results from something not of the player's doing (such as a random roll).

Earlier, Jay R lumped several changes in a game state as "cosmetic", when what they really are is narrative: even if you're locked into an immortal hero's journey, the question "what sort of a hero are you?" is obviously meaningful.

Meanwhile, NichG is largely pointing to a vast body of gameplay elements that can serve multiple aesthetics, challenge included, without there being a need for threat of elimination. Base building, or any building, is a great example: even if no-one gets eliminated for not building an impressive thing, the fact that some things are harder to build than others creates potential for test of skill and the observation that someone else might be building a more impressive thing than you creates potential for competition.

[kyoryu]

What I'm pushing back against here is the tendency to focus on 'delta' coming from things like combats, failures, etc. There's nothing wrong with having character death be a thing or failures be a thing, but its like spiciness (heat) in cooking - if your idea of how to make meals tasty was too strongly focused on 'what is the correct spice level to use?' you're really limiting yourself.

Without any failure, risk, death, combat, etc, you can still have lots of meaningful choices. The richest, most meaningful choices actually tend to occur outside of those sorts of conflicts, because 'what you want' is pretty clear once initiative has been rolled - mostly you want to kill the enemies, not die, and use as few permanent resources as possible. The consequence for screwing up is strong (e.g. it has an inherently very high saliency, like spicy food getting your attention), but outside of that its more one-note than other things.

Compare that with a choice like 'what do we do with this old castle we cleared out near the border town?'. Turn it into a base for the PCs? Okay, how would you like your base, what sorts of features do you want it to have, etc? Not just aesthetic - you'll be using those resources for the rest of the campaign perhaps. Turn it over to the local nobility in exchange for favors? Okay, interesting, which noble house in particular do you want to make stronger and more influential, and what kind of favors are you looking for? Just leave it sitting there abandoned? Okay, lets figure out what sorts of things might move back into it over the next year or two, that will probably have some consequences for the region. Break it down for raw materials and sell them? Okay, that takes some time and becomes a local endeavor, so this is what we're doing for the next few sessions - interactions with the local Masons' guild, Merchants' guild, finding buyers, finding ways to turn the new influx of wealth into stuff you actually want (and deciding what you actually want!), etc.

Those choices don't have the sting of losing progress when you make them 'wrong', but they can be very complex and have ongoing effects and can make lots of different things about the world matter. If combat and death is the spice in this metaphor, those sorts of 'what do we want?' choices are the umami, the aromatics, etc.

I agree that there are lots of interesting choices outside of failure that need to be leaned on, but I'd like to point out that there's lots of interesting failures that can happen that don't need to involve death or even "loss of progress", but can instead look more like the "fork in the road" decisions you're talking about, with the only real difference being that one of the possible outcomes is more preferred.

In fact, that's my general way that I run almost any kind of scene. I've referred to this as basically whether you're using a "gated challenge" type structure, or a "fork in the road" structure. I vastly prefer the fork-in-the-road. It's less "you lost" and more "you didn't get what you wanted".

[gbaji]

I don't know that high lethality is bad for Delta V (getting beyond the "failure doesn't have to mean death" discussion). What's important is that failure/death be, at some level, the result of decisions made, and preferably the result of multiple decisions made.

If you make a decision to charge a group of enemies by yourself, get beat up, refuse to retreat, and finally get killed? I find that perfectly in line with Delta V. If you're walking along in "cut scene time", there's an ambush that you had no ability to avoid, you roll for surprise and lose, the enemy takes their shot and kills you, that absolutely screws with Delta V. There's a result there that was not really a result of any decision you made.

Pretty much exactly this. I have a general rule as a GM that I will not kill a PC as a result of a random die roll that the PC could not avoid taking in the first place. Note, howevever, that this doesn't necessarily mean that at the moment, they could choose a different action/rolll (though sometimes it can!), but that the condtions in which there will ever be a die roll that may result in a character death (or, well, any significantly negative impact to a character) will only occur as a result of a series of previously made decisions by the players collectively, if not that player specifically.

You have to have made the decision to put your character into the risky situation in the first place, knowing at least in general what those risks were, and typically having the opportunity to use time and skills to gain even more info about said risks, to get into the situation in the first place. That's also not to say that a monumentally poor series of die rolls might not also land your character into a dead state, but that's extremely rare to even come up without some decisions going on ahead of time. I just tend to prefer to allow the players to have significant agency in terms of the risk they take with their characters. Of course, since they're typically playing in scenarios where there are risks that must be taken to obtain some reward or outcome that they desire, that's kinda part of the game. But the players know this going into it, and make those decisions themselves.

The time I died when I got mobbed by Goblins and stabbed to death in session 1? Just annoying that I bothered to put anytime in making a character beyond the numbers in the first place. The time I made a character and got killed on my first die roll of a d6 Star Wars game, just annoying. Dying in a long running campaign thanks to a "Save or Die" in the middle of the campaign, just annoying as I was the only one going through "new player" mode in a well-established group. In those situations, I learned to not care about my character as they were just an expendable game token, no more, no less. Who cares if I walk face first into a death trap? No one. If I LeRoy Jenkins a dragon by myself? No one. Just make a new token! Boring.

That's kind of the trick though, isn't it? Death (or equivalent) has to be enough of a risk for the character to care to try to avoid it. But death cannot be such a high risk that it feels inevitable, so there's no point in caring either. Either extreme leads to players who just don't care about the actions and outcomes of their characters.

The trick is finding the correct balance between those two IMO.


On the broader topic of decisions and outcomes though, my opinion is that you do want the PC choices to matter. But it's kinda like the death thing. If the PC choices have too much effect on the outcome, then they lose meaning. If me deciding "I want to do X" results in the GM creating the conditions for X to occur, obtaining X outcome doesn't feel very satisfying. On the flip side, if all paths lead to X, that's not so great either. There are a number of ways to manage this. My personal method is to not "force outcome X", but merely present X as "an outcome that may occur". And yes, sometimes this is a bit contrived. If I've written a scenario, with a bbeg who is doing something evil, and the method to defeat said bbeg is to figure out what he's doing, then find the <whatever> needed to stop him, which in turn requires going to <various places> to obtain that, then going to <some other place> to defeate said bbeg's plot, then those things are somewhat set. But the "how" is not. Where they go to get the info to figure things out may vary wildly (and I"m absolutely not above PC cleverness and out of the box thiinking playing hugely into this). How they go about getting to various locations to get <whatever it is they need> may also vary (and ofen the order, or even the specific destinations may vary as well, depending on the specifics). And, of course, the PCs are free to choose other things along the way, go off on tangents, etc.

My rule on this is "what are the players interested in?". If the players are enjoying the adventure, and happily moving along, then I'll just keep putting clues in, and letting them follow them to "X". If they decide they want to do something else for a while, then they can go do that. But the bbeg, and the "X" doesn't disappear just because of that. The point is that where and when they engage in the "main plot" (for lack of a better term) is 100% up to them. And honestly, I pretty much never even have any issues with this. It's pretty rare for players to go "well, we ran into some clues about this evil plot going on, and we have some information that we can follow up on if we want, but hey... let's just ignore that and go roaming around the countryside randomly looking for stuff instead". Not for more than minor diviersions that is. But that's part of the worldbuilding as well. I pretty much always put in some other things that may be going on in the same time/location that they are in. Usually minor things, generally not leading to anything major. But there, nonetheless. So if they find themselves in <some city> tracking down someone they need to interact with in some way for the adventure they're on, there will often be other things going on as well in that city. Some may just be random type encounters. Some may be small mini adventures.

The point is that they get to chart their course though the stuff I've placed in the world around them. And yes, I'll create content (on the fly sometimes even!), but usually only as a result of character actions. If they choose to do something, there will be responses from the world and the elements within it, to those actions. And sometimes, this may lead them into some interesting situations. But I tend to not go too far with that. As I said above, I don't put things in because a player wants it. A player can certainly express a desire to achieve some outcome, but if they want it, they need to come to me with the actions their character is doing that lead to that outcome. I'm not just going to drop it in front of them.


I guess the interesting question here is where the distinction between "actual delta in outcomes" and "percieved delta in actions" is, and which is actually more important. To me, at the end of the day, one of the primary objectives as a GM is to create an enjoyable game for the players. Too much "gimme", and they wont enjoy it. Too much "gotcha", and they wont either. I find that if you can find the right balance of "I'm going to put things into the world" and "you get to decide what to do with/about those things", the players tend to enjoy that a lot. And yeah, not surprisingly, a lot of the time that is about writing a pretty straightfoward adventure, with a relatively concrete set of "steps/chapters" within it, that the PCs are generally going to mostly just follow along. But as long as there are enough interesting bits in there, and details, and choices along the way that affect how other things play out (even if they may actually be somewhat ornamental in the grand scheme of things), that's plenty of "we have choices and they matter" for most players.

Also helps that I often get ideas for scenarios based on those "side choices/actions" PCs made along the way in some past adventure. Heck. I'd say that somewhere around 50% or more of my adventure ideas are created by looking at "loose ends" from some previous action or event that the PCs were involved in at some point in the past.

[KorvinStarmast]

It's two different monsters so far, but you (or someone else) could repeat the same "what about this other monster?" until we're all sick of it. I'm telling to stop now because it doesn't matter. It doesn't even matter for the argument you're supposedly countering: if someone's of the opinion "instant death monsters should be signposted", then their answer to any "what about this monster that isn't signposted?" would presumably be "then add the signpost or don't use the monster". Nothing forces anyone to use any monster in a completely unpredictable manner.

Speaking of "insta death monsters" The Butcher in Original Diablo was a great example. (I ended up with 3 different ways of dealing with him, and then stumbled on a fourth).

[MoiMagnus]

That's kind of the trick though, isn't it? Death (or equivalent) has to be enough of a risk for the character to care to try to avoid it. But death cannot be such a high risk that it feels inevitable, so there's no point in caring either. Either extreme leads to players who just don't care about the actions and outcomes of their characters.

The trick is finding the correct balance between those two IMO.

And because of randomness, the balance is actually quite hard to reach, because some players will be unlucky.

Among the D&D players, is it acceptable that 10% of them have a sequence of bad luck and stop caring about their character after multiple pointless deaths in a row? And 1%? What is the % of players you are ready to sacrifice on the altar for the other players to feel like there was a risk in first place?

Admittedly, if you always play with the same group of friend, the GM can actually fix this by toning up or down the mortality of your games on-the-fly, depending on what the players want and depending on what happened previously (taking feedback from your players about "did that death feel fair/ok?" or "did the enemies feel threatening/dangerous?").

But I fully understand why D&D game designers progressively pushed the game toward being less and less deadly, especially at low level.

[Vahnavoi]

@MoiMagnus: humans demonstrably care about, even get addicted to, games of chance where they have less than 50% chance of victory. So your argument for why chance of death has been lowered in later editions of D&D is incomplete at best, and fallacious at worst. There isn't such a straightforward relationship between frequency of loss and how much a player cares about their participation in a game.

The actual reason chance of death has been lowered, is more likely that later developers (and players) idealize the concept of taking the same character through an entire campaign, and that is impossible if random death is too common. They may presume that players remain more invested this way but, again, the relationship is not so straightforward that this ought to be taken for granted.

[gbaji]

And because of randomness, the balance is actually quite hard to reach, because some players will be unlucky.

Among the D&D players, is it acceptable that 10% of them have a sequence of bad luck and stop caring about their character after multiple pointless deaths in a row? And 1%? What is the % of players you are ready to sacrifice on the altar for the other players to feel like there was a risk in first place?

I think those aren't necessarily connected though. Especially the "mulltiple pointless deaths" bit. If you are following the rules I laid out, there should never be *any* "pointless deaths".

If you find that your players (or some of them) are upset at multiple pointless deaths, the correct answer isn't to sacrifice the enjoyment of the rest of the players by putting the equivalent of fluffy child protection systems in place, but to change the nature of the things you are putting in your scenarios which are leading to these deaths. Characters should only be in a position where die roles may result in their deaths, if they have previously made a decision that put themselves in that position, presumably because they judged the objectives they are trying to achieve to be "worth the risk". That can (should) never make that death pointless. If they are fighting the invading monsters terrorizing the local village, and one of their members dies heroically because they choose to run off to engage some of the monsters who were about to chow down on one of the village children, then that death absolutely wasn't pointless.

If, on the other hand, the same character is walking along, and is hit by a poisoned assassins dart out of the blue, fails their save, and dies? That's pointless. If they are exploring some random corridor, fall into a spike trap, and die? That is pointless (you may get the pattern here and why I'm not a huge fan of "make a roll or die" style traps/attacks). Change your adventures to have more of the former type of "these are situations where my character may die", and fewer (maybe even "none") of the latter.

It's actually not that hard at all to tailor adventures to this rule. And, it actually usually makes a lot more sense from a realistic point of view (seriously, stop and think about what kind of traps or defenses you'd realistically want to have inside your own home or work area). Unfortunately, D&D has a long long history of "save or die" mechanics, and it can be tough to fight against the trend. But once you reject stuff like that, you'll find that your games not only run better, and your adventures make more thematic sense, but your players will enjoy them a heck of a lot more.

But I fully understand why D&D game designers progressively pushed the game toward being less and less deadly, especially at low level.

Yeah. Unfortunately, in many cases, instead of just tossing the "save or <something bad>" mechanic, often they just fluffify the <something bad> instead.

I'm saying "don't do that at all". That's not to say you can't have poisons, and can't say "poisons do <some damage/stat effect> rather than death" (that's actually a very good modification to the rules). And it's not to say that you can't have powerful spell effects that may incapacitate characters from time to time (but not kill them). But those things should be side events to the "main conflict/resolution" process. Give the PCs goals/objectives, and then obstacles to overcome to achieve them. Make overcoming those obstacles include reasonable risks. But maybe try to make those risks a bit more granular in nature. Combat is a great example of this, since it's often quite rare that someone just straight up dies in one round in combat (can happen, but usually does not).

It's really surprising how often this comes up when talking about D&D, given the HP inflation methodology of that game, and the more or less "resource depletion" nature of encounters. Deaths to characters in D&D should very very rarely occur except in a case where they're in a long tough fight, low on HPs, low on spells/abilities, and the PCs make the decision to "keep on pushing" to win. And, unless the combat and resolution thereoff is also pointless (which maybe says something about the scenario design), if a death occurs in that situation, it should be highly heroic, and not at all pointless.

And just to loop back to the point I made earlier about "give the PC goals/objectives". That's critically important. I've found (after decades of GMing) that players vastly prefer to be the ones "driving the plot", rather than the other way around. Adventures just plain work better if it's the PCs who learn about the evil plot/scheme/monsters/whatever that are about, and they decide on a course of action to deal with them, than if they are simply on the receiving end of attacks from said evil things. The former method puts the PCs "in the know" about the opponents, and often their greatest defense is that said evil things don't actually know (yet) who they are. They are on the offensive. They are the ones raiding the bbegs minion's locations. They're the ones finding journals, business records, shipping manifests, or whatever that leads them to the next stage of the adventure. They're the ones who are "surprising" the bad guys (most of the time).

When adventures are run the other way around, and it's the bad guys on the offensive (targeting the PCs), and the PCs on the defensive, that's when it can actually be very very hard to rationalize why the bad guys aren't using any of a host of incredibly simple and deadly things against the PCs. And this is what can lead to the players possibly feeling their deaths were pointless. So... always make your adventures such that, baring epic failure on their part, the directionality of the adventure and most encounters is the other way around. Not only will that dramatically decrease the death rate of the PCs, but it will *also* make the players feel like they are more actively driving the adventure rather than the other way around. It's their choices that lead to most encounters. They choose the when/where. They will be massively more engaged in the story/plot and events occuring around them if you make that actually be how they succeed. If they are merely passively waiting for the GM to attack them with monsters, then they wont have that investment, and are much more likely to feel that any deaths that do occure are actually pointless.


Oh. Funny followup to a point I made earlier in the thread, about my players and how they will pretty much always choose to follow threads I lay in front of them. I'm currently the running GM in our campaign. I told the other main GM that I was going to run a shortish "local adventure" (one character each), and then would run a longer/tougher adventure (two each). I had an idea for the short one, and was basically percolating ideas for the longer one. But, while writing the short one, there were basically two threats they had to deal with (all revolving around a single small town in the kingdom they were in). One threat was pretty much a basic "band of bandits in the hills". The other was a more mysterious threat that a local knight fell to. The party clues into this when tracking down the missing knight (who was investigating the bandits, but they rapidly discover he also started tracking down a series of mysterious disappearances that had been going on, quietly, for a decade or so). Turned out it was a pack of ghouls, stealing corpses from the graveyard nearby, and occasionally snacking on indigent folks (and sometimes making more ghouls). Great. They deal with that. But, of course, nothing is so straight. In our game, there are actual just normal "ghouls" in the game, but "ghoul" is also a term used to apply to human servants of vampire, who have been granted blood boons (yeah, got that term from V:tM, right). Said ghouls are just like normal folks, enhanced a bit, and with extended life, as long as their vampire master regularly provides them with blood. If not... they will eventually devolve into something more like the standard ghouls (including the ability to turn victims into actual normal ghouls as well).

As it turns out, like 20-25 years back, I ran a vampire adventure (out of the same small town, which none of the players actually remembered specifically). Said vampire was part of a larger plot going on at the time, and had kinda struck out on his own after some other even earlier events. Part of his plot was a terrible stereotype. He falls in love with one of the PCs, and is seducing her, under the guise of a friendly gem merchant. Players eventually figure out what's going on, have to rush to find his lair (he has a small cave complex in the nearby hills where he does his evil stuff, cause he like entertains guests in his actual manor home in town), and save her before he can turn her into a vampire. He escapes, but not before stealing some rare alchemetical components related to some work he'd been asking the local alchemist to do for him (cause he was such a friend to everyone, right?). Lots more to the actual scenario, but he had a foot/doorman and a butler. And no one ever found out what happened to them... Well, they were abandoned when their master had to leave suddenly, and now they're reduced to hiding in the secret basement under the old manor house, and have developed over time a need to feed on sentient flesh. And have built up a small group of ghouls, and are "awating the return of our master". And yeah, it really ratched up the paranoia of the group when they entered the basement (from a tunnel that had been dug to one of the crypts in the graveyard), and while in the middle of the fight notice a "really ornate coffin" in the one clean area of the room.

Here's where it gets really relevant. When they first defeated this vampire, they found a bookcase in his lair with a bunch of journals. From them, they learned how/why he and his vampire friends had arrived in the area in the first place, and what they were doing (which tied into a larger plot going on). And it contained some broad info about what he was doing specifically (and why he was interested in the alchemetical stuff). But the latest journal was missing. They assumed he took it with him or something. Well, in the basement with the ghouls, in addition to boxes of decently valuable gems, and some spell books, was... the last journal. With the stuff he was doing/planning right when they first encountered him. So yeah. I seriously did no prompting at all. They were immediately like "Ok. We're assembling a group and figuring out where he went and going after him!".

So. Now, that's the thrust of the longer adventure. I'll certainly pull a few elements from other things I had planned, but it'll be "figure out what the vampire was planning, follow his tracks, and take him down". I had a different adventure they could have gone on, but this is where they want to go, so that's what I'm going to run for them. So... loose end from one adventure (20+ years old!) turns into adventure today. And they could certainly have said "well, that's interesting info, and we'll maybe follow up on it someday, but... what adventure do you have for us now?", and I would have just run another adventure idea I had instead. Fortunately for them, vampire (well, ones that live a long time) tend to make long range plans, and take their time setting things up, so the journal will be useful to them. At least to a point (can't make things too easy on them, can I?).

I guess what I'm illustrating is that you absolutely can run somewhat linear individual adventures, but in a more free flowing open campaign. Just let your players decide what clues they follow up on, and how they go about it. Doesn't mean I'm not going to plop something right in front of them, if they're not actively doing something else (which was what I was planning to do initially anyway), but I've also found that by picking up old plot lines, or re-awakening past events and making them relevant, the players will practically leap into action (at least that's how my table works). I'm rarely sitting there with a table full of players looking at me, waiting for me to tell them what happens next. I set the wheels in motion, and then let them drive the cart.

[kyoryu]

But I fully understand why D&D game designers progressively pushed the game toward being less and less deadly, especially at low level.

I think it was just two things, really.

1. The switch to presuming a single character vs a stable of characters.
2. The addition of long-term story, and the difficulties that presented with narrative continuity.

And, really, I think it's mostly the first, though that was probably prompted by the second.

[Saintheart]

If you want rolling the dice to be meaningful, i.e. have a potential impact on the outcome, then you have to be able to win or lose by luck. Saying you shouldn't ever lose because of bad luck is equivalent to saying you should never win by luck and also that you shouldn't have luck at all, at which point you're better off playing chess. Basically if you include meaningful dice, you're making the meaningful choice to win or lose randomly.

Lots of good thoughts here that honestly I haven't read through thoroughly, but this seems like a good entry point, especially from some of the immediately preceding posts which basically come down to: 'you can more readily handle character death if your chose to do so'.

The above quote I think is really important, but I'm going to start by disagreeing with it: I don't think dice are part of the game in order to make the game meaningful. The dice are there to represent (I won't say simulate) an element of reality that hits certain sections of our meaning-seeking brains.

Dice, and I admit on this one I've been persuaded by the Angry GM's blogs on it, are present in a RPG to represent cruel reality that doesn't care about your dreams, intentions, etc. Assuming a fair dice, there is a remorseless chance of your stated intent going wrong when it's applied to reality. Under D&D, in combat contexts especially, this becomes an all-but-guaranteed chance of things going wrong in the natural 1. The stripped-down, repeating 'core program' of a RPG in progress is:
1) DM describes situation
2) Asks what characters do
3) Players specify what their characters do, or attempt to do
4) DM adjudicates the results of that attempt.

Step (4) is the only point at which dice come into the picture, and only then if the DM feels a dice roll is called for because the outcome is not certain. If you knew the player would fail, as a DM, you'd tell them so, and if you knew what they wanted to do would succeed, as a DM, you wouldn't call for the roll.

The dice give reality's opacity, cruel objectivity, and uncertainty their say on our intentions. A say that can't really be argued with. A say that might result in our deaths depending on the context. Reality is arbitrary if not capricious and it's often unfair (at least as we perceive fairness). Or at the very least there are thousands of elements operating on our day-to-day existence that we can't account for or calculate for or prepare for. Our life in the world is struggle against the arbitrariness and opacity of reality. In a RPG, that struggle is symbolically represented by the pluses and minuses to our stats. The modifiers allow us to kid ourselves - in a good and sane way, just as we do in our normal lives - that our actions have some sort of influence on reality even if they can't necessarily affect the outcome. That's also why the natural 1 and the natural 20 are in place -- they also remind us that, no matter how big our modifier is, things can go unexpectedly south on us, and there's nothing our accumulated experience or intellect can do about it.

The dice in my view don't represent meaningful choice. The main reason for their existence is to sow the RPG experience with a 'cue' that helps us convince ourselves it's more real than it first appears to be. Remove the dice, and as said you've either got a zero-sum, fixed-outcome game like chess, or you have freeform RPGing, which is much more about wish fulfilment and (not coincidentally) often has big problems resolving things like fights. Or you have a 'god game' RPG like Microscope where the experience being simulated is the construction of a narrative, not a representation of reality as a RPG tries to be.

That said, there is some truth that having dice means you're making a choice to win or lose by luck. My experience tends to be that losing by luck is a little more tolerable if some choice was involved into getting into the losing situation. Like that failed Str check mentioned above, or where the party is warned that X course is dangerous and they choose to take the odds. I've even experimented with quietly taking a player aside when his HP hits -10 (before advising publicly) and asking whether he wants to continue the character or not. This doesn't duplicate reality, but then, a RPG isn't meant to duplicate reality any more than a painting or a novel produces an exact facsimile of a moment or moments in time. They are all there to induce an emotional response, an emotional kick, and teach us in some small, imperfect way, about how to better handle our lives.

[Herbert_W]

If you want rolling the dice to be meaningful, i.e. have a potential impact on the outcome, then you have to be able to win or lose by luck. Saying you shouldn't ever lose because of bad luck is equivalent to saying you should never win by luck and also that you shouldn't have luck at all, at which point you're better off playing chess. Basically if you include meaningful dice, you're making the meaningful choice to win or lose randomly.

There's a hidden assumption here that I'd like to call into question: being able to win or loose by luck is not the same thing as always being able to win or loose by luck.

Success or failure aren't binary states. There's a sliding scale:

• Win, which usually means dead enemies, loot, and XP.
• Draw, which could be you escaping from your enemies or vice versa.
• Loose, meaning story failure, loss of items, level drain, or character death.

Each of these points can be expanded into a finer scale based on resource expenditure. There's some weeds that we could get lost in here. You could make the scale even finer by further subdivision (how easily can the lost resources be recovered?) or acknowledging incomparibles (save an NPC vs. gain a better sword).

For simplicity, let's just suppose that you have a resource budget for the day/adventure/arc/whatever and each encounter could put you (1) over, (2) on, or (3) under that budget.

Multiplying this out, you have nine meaningfully distinct points on the scale of success. Again, there's weeds that you could get lost in here. I'm not saying that there's exactly nine meaningfully distinct points, just that there's a bunch of 'em.

So, what moves you up or down this scale? Well, there's:

• Player choices
• DM choices
• Luck

You might think of these as being like a pie chart. The more significant any one is, the less significant the other two can be.

For a game to respect player agency, luck and the DM should both have a small enough role in deciding outcomes to allow room for choice to mater. We already know that railroading is widely considered to be bad: a DM who seizes 100% control over outcomes takes away the players' feeling of agency. (This applies whether the DM railroads to success or failure. The latter may be more frustrating, but the former is still unsatisfying.) Granting luck the same overwhelming power over outcomes would create the same problem for the same reason.

Yet, as others have pointed out, at least one of these needs to have a large enough influence to be felt for a game to feel like a real world with real challenges.

This creates a tradeoff. Luck (and equally DM guidance, but we're primarily talking about luck here) having too much or too little influence both cause problems. There's a happy medium somewhere in between these extremes. The optimal point will vary based on player preference and the genre of game, but it's there.

If we make a binary choice between extremes, both options suck. We want to find the happy medium.