D&D 5E Failing saves is...ok?

[Tony Vargas]

Well, I think there are two separate issues, as I outlined above. One is the "impossible save" vs. the "auto-succeed on a 20 (and auto-fail on a 1)". The other is the whole scheme, wherein some saves will just never be as good as others despite leveling up.

The former issue is a minor one. BA should generally avoid the extreme cases of failure on a 20 or success on a 1, and the DM should usually avoid them, as well, by simply not calling for a roll.

The latter is misstated, IMHO. It's not that some saves will never be as good as others, in spite of leveling. That's inevitable with stats adding to saves and straight-20s on all six stats being un-achievable. Rather, it's that some saves will stay exactly the same, for a given character, in spite of leveling from 1-20, since most saves for most classes get no proficiency bonus, and you can't boost /all/ 6 stats with only 4 ASIs, while save-forcing mechanics, over the same level, will scale dramatically, because they do use both typically-bast-stat for the class /and/ proficiency (or the equivalent for monsters).

That's not just be better at some saves than others, that's not just being bad at some saves, that's getting steadily worse at standing up to the challenges you face as you level.

That's in stark contrast to classic game. And, it doesn't have to be, because 5e /has/ found ways of handling scaling with some mechanics. Take Sleep, for instance. In the classic game, Sleep affected a random number of targets based on their HD/level, over 4, you were unaffected, otherwise, no save. In 5e, Sleep compares a hp roll to remaining hps and affects targets it reduces to 0. Because hps baloon with level, you rapidly become very resistant to Sleep. It's not as absolute in either case as it was back in the day, but it's still following the same basic dynamic. Very effective vs low-level targets, far less so (though not completely useless as it was back then) as targets get higher level.

All right. It's uncontroversial (AFAIK) to say that high-level 5E PCs will fail saves more often than high-level AD&D PCs.

It would be controversial to say that high-level 5E PCs suffer more from failed saves than high-level AD&D PCs.

It's a more nuanced statement, and there's clearly some resistance to it, but I don't think it's that important compared to the fact of failing saves more often.

Still, OT1H, 5e PCs will suffer more from failed saves, because 5e calls for at least as many saves as prior editions ever did, and they will fail more of them, but, OTOH...

. The consequences to failed saves in 5E are pretty mild.

5e PCs probably suffer less as a consequence of failing a given individual save against the same thing.

Take Hold Person. Being Held is about as bad in 5e as in most other editions, but in the classic game, you got only the one save, while starting in 3.5, you got repeated saves to reduce the duration. When 3.5 introduced that, there was some controversy, because, the theory went, the repeated saves made you a target for Coup de Grace, while the one-and-done failed save made you a non-factor in the balance of the combat, who would probably be left alone, therefor being able to repeat the save was 'worse.' Whether that logic makes sense to you, it's how the spell has worked ever sense (more or less, 'save ends' duration was different in 4e, not being directly based on the power of either the attacker nor defender, just a random duration that was 55% likely to end each round). If you're not instantly-prioritized for being able to make the save again (or if you are prioritized even though no further saves are likely forthcoming), though, then being able to save again is strictly better than no. Even if you need a natural 20. However, being able to save on 6 (or more likely, since the classic game assumed magic items & 5e does not, less, even a 2), is probably better than getting a chance to save each round, but only making it on a 20. Well, there'd be an inflection point after some number of rounds, anyway...

So, the consequences are arguably marginally less severe in that case, in 5e. In plenty of other cases - save or suffer the poisoned condition vs save or die, OTOH, saves in 5e have much less severe consequences.

However, I think the line where the two considerations - chance of failing vs consequences of failing - would 'balance' and failing saves more often as you leveled instead of less often would be acceptable, would be when all consequences are hp-denominated, and the scaling of such 'damage' was slower than the scaling of maximum hps. In other words, if everything calling for a save worked like Disintegrate or even Sleep.

That's not how saves work in 5e:

So after the umpteenth iteration of "5e effects are mild" and "failing saves is interesting and fun", I decided to take a closer look at what monsters actually do to you.

Aboleth, Banshee, Cambion, Carrion Crawler, Chuul, Cockatrice, Dryad, Spellcasting Enemies, ...
... I only got as far as "E" in the Monster Manual before I stopped caring.

So, if those few monsters (and their F-Z comrades), and spells in general, were re-written to have save effects also gated by comparison to current hps, then saves being as likely to fail as attacks are to hit would start to be reasonable.

Of course, because there are 6 saves and only 1 AC, it's a lot harder to get saves and attacks vs AC on about the same footing.

 

[FormerlyHemlock]

It's a more nuanced statement, and there's clearly some resistance to it, but I don't think it's that important compared to the fact of failing saves more often.

Clearly you don't--you've said so repeatedly in this thread. That's personal taste however; there are others who couldn't care less if saves get failed often. E.g. who cares if you fail 95% of your saves vs. Aboleth charm if you've got a Paladin of Devotion fighting next to you and therefore can't be charmed? 95% of zero is still zero. (Ditto if you're wearing a Protection From Evil spell.)

5E's designers have said this is the way it's supposed to work.

we make no assumptions on the DM's side of the game that the player's attack and spell accuracy, or their defenses, increase as a result of gaining levels. Instead, we represent the difference in characters of various levels primarily through their hit points, the amount of damage they deal, and the various new abilities they have gained.

Like it or not, that's how Bounded Accuracy works, and how 5E was designed. In practice it seems to work out just fine. Counterexamples are welcome but so far none have been provided.

However, I think the line where the two considerations - chance of failing vs consequences of failing - would 'balance' and failing saves more often as you leveled instead of less often would be acceptable, would be when all consequences are hp-denominated, and the scaling of such 'damage' was slower than the scaling of maximum hps. In other words, if everything calling for a save worked like Disintegrate or even Sleep.

That's not how saves work in 5e

I understand your point, but I disagree that such a change is needed. There are already a ton of mitigations for failed saves in 5E, and most failed saves don't cripple you anyway. (Even DEATH usually doesn't hurt you that much in 5E.) There are a handful of really bad saves to fail, but most of this thread has been spent discussing things like the Frightened condition which aren't severely debilitating even if you do fail the save, and has plenty of mitigations.

I understand that it bothers some people on this thread for high-level fighters to still slip on DC 10 ball bearings spilled by a kobold, or for certain high-level wizards to always take full damage from dragon breath; I don't really grok why it bothers you folks, but there's no accounting for taste so whatever. To everyone else though, all the suggestions advanced on this thread look like solutions in search of a problem.

Notably, no one except James has ever attempted to lay out an actual problem caused by the way saving throws work in 5E. The only example you gave in your most recent post, Tony, was Hold Person, and you were using it to illustrate how 5E's consequences are less severe than they used to be.

 

[Tony Vargas]

Like it or not, that's how Bounded Accuracy works, and how 5E was designed. In practice it seems to work out just fine. Counterexamples are welcome but so far none have been provided.

Saves are a counter-example, or, rather, save DCs are an example of BA being routinely broken.

I understand your point, but I disagree that such a change is needed.

That's fair enough. The point isn't that it's wrong for 5e to favor anything (monster, NPC, or PC) that forces saves rather than makes attacks, nor to overwhelmingly favor the sub-set of those who can force a selection of saves and thus hammer enemies' weakest saves. Rather, it's a matter of whether we individually deem that innate imbalance undesirable, and how we, as DMs, might fix it in the context of a campaign. I mean, that's prettymuch the context of any discussion, because the power to make any changes rests with the DM. I know they often get couched as blanket changes that WotC must make - and force on everyone - to save the game from itself, but that's not the case, anymore (in the edition war years, it was very much the case, and 5e is the result - both in the way it's default form slavishly caters to 'h4ter' prejudices, and in the way it's designed to be faux-'modular' and DM-empowering, so each individual table can use it as a starting point and bang it into the shape they want - there's just a lot more banging required in some cases than others).

So, one possible place to kick the system if it's not working in this particular way is in the 'bad' saving throw category. Give proficiency to saves across the board, or give half-proficiency or proficiency-2 to 'bad' saves. It shores up the low end of save bonuses, and keeps Bounded Accuracy a little more bounded.

Actually reaching for the save dynamic of 1e, which I often hold up as a contrast, would be a more involved project...

But one small stab at it might be a feat, something like:

Magic Resistance
Preq: You must have no spell-casting or other magical or supernatural abilities.
Benefit: When you are affected by a supernatural ability, roll a saving throw using your best save bonus against the DC of the ability (even if it doesn't normally have one, the DM will determine a DC based on the creature using the ability). On a success, the ability has no effect on you. On a failure, it is resolved normally, including any saving throws.
If you wish to be affected voluntarily, you must still make the save, but as an unmodified d20 check.

I don't really grok why it bothers you folks, but there's no accounting for taste so whatever.

Would it bother you if D&D characters never gained HD after first level? It's like that.

Notably, no one except James has ever attempted to lay out an actual problem caused by the way saving throws work in 5E. The only example you gave in your most recent post, Tony, was Hold Person, and you were using it to illustrate how 5E's consequences are less severe than they used to be.

That may be how you saw it, that's not how it was intended. Rather, I feel it illustrated how even when the consequence seems like it must be less severe (save every round to end is certainly less severe than save once and done!), the fact that saves can be so much harder to make the save can cause it to be more severe.

And, I didn't see a need to repeat James's example, nor to look up their F-Z companions. There are plenty of cases of saving throws having severe, hp-bypassing consequences. That breaks the idea of scaling primarily with hps/damage that supposedly excuses breaks in BA.

 

[FormerlyHemlock]

Saves are a counter-example, or, rather, save DCs are an example of BA being routinely broken.

"Saves" isn't a counterexample, it's just a concept in a vacuum. If you can provide an example of a case where BA leads to a broken game due to saves not advancing, we can discuss that case. But simply re-iterating that saves don't increase with level isn't illustrating a problem with BA--it's merely re-iterating one of BA's design principles.

Would it bother you if D&D characters never gained HD after first level? It's like that.

In that case it would be trivial to illustrate my dissatisfaction with an example. I could just say, "four 8th level PCs of the iconic classes can't stand up to a couple of CR 5 Flameskulls, which seems wrong to me. Furthermore, PCs wind up playing by completely different rules than the monsters do. Even a low-CR 'Berserker' monster, which is notionally supposed to represent a low-fidelity NPC somewhat like a Barbarian, has an order of magnitude more HP than a 10th level Berserker Barbarian. This seems wrong to me on multiple levels. PCs can't handle monsters of the difficulty they're supposed to be able to handle; they're too vulnerable to friendly fire like Fireballs, which consequently makes Evokers' Sculpt Spell more valuable than I think it should be. Finally, if I want to play a game without HP advancement, I'll play GURPS."

Anyway, if you're just trying to have a discussion with your fellow DMs who share your distaste for Bounded Accuracy as implemented, I'd best bow out of the discussion, since I'm fine with that part of 5E.

 

[Tony Vargas]

Anyway, if you're just trying to have a discussion with your fellow DMs who share your distaste for Bounded Accuracy as implemented, I'd best bow out of the discussion, since I'm fine with that part of 5E.

BA isn't perfectly-implemented (what can be, in the 5e philosophy of standard-rules-as-starting-point?), but the same issue ("saves"* net getting worse as you leveled) has existed for a while: even more dramatically in 3e, in which accuracy was radically unbounded, and to an extent, in 4e, where it was rapidly-but-evenly scaling. It's been a major change to the game's dynamics that's rarely addressed - like the trend of removing limitations from casting with every single edition, or the 5MWD....











* scare quotes necessitated by 4e not technically using 'save' the same way, but having 'non-AC' defenses - darn that non-conforming red-headed step-child of an edition. ;P

 

[FormerlyHemlock]

BA isn't perfectly-implemented (what can be, in the 5e philosophy of standard-rules-as-starting-point?), but the same issue ("saves"* net getting worse as you leveled) has existed for a while: even more dramatically in 3e, in which accuracy was radically unbounded, and to an extent, in 4e, where it was rapidly-but-evenly scaling. It's been a major change to the game's dynamics that's rarely addressed - like the trend of removing limitations from casting with every single edition, or the 5MWD....

It's not an implementation issue with BA--it's part of the definition of BA, according to Rodney Thompson, the WotC designer who coined the term "Bounded Accuracy" in the first place. I've quoted his words at least twice in this thread. PCs advance by getting more powers (and HP and damage), not necessarily larger pluses to their d20 rolls. If a wizard's Dex save vs. Meteor Swarm is crummy, so he Counterspells the Meteor Swarm instead of saving against it--that's bounded accuracy in its purest form right there.

 

[Tony Vargas]

It's not an implementation issue with BA--it's part of the definition of BA, according to Rodney Thompson, the WotC designer who coined the term "Bounded Accuracy" in the first place. I've quoted his words at least twice in this thread.

Nod, I re-quoted it, because it spoke to the way the issue manifests in 5e. Scaling manifests much more in hps/damage in 5e than in bonuses/DCs, so when a hps are bypassed there's a potential problem. There are fewer saves that bypass hps in 5e, but there are still quite enough to be a problem.

If a wizard's Dex save vs. Meteor Swarm is crummy, so he Counterspells the Meteor Swarm instead of saving against it--that's bounded accuracy in its purest form right there.

I don't see how it can be: The exact same thing could have happened in 3.x, in which accuracy was wildly un-bounded, for instance.

I think part of the problem we're having is seeing only the bonus side of the equation as 'bounded,' and perhaps, also in seeing only the upper end of it as bounded. AC and save DCs also need to be bounded, for instance, and bounded relative to plausible bonuses used to hit them. Attack bonuses are typically an attacker's best bonus. They pick the weapon that sues their STR or DEX, whichever is higher and in which they're proficient or use a cantrip/spell which'll generally be off their best stat and always include proficiency. Even really weak fractional-CR critters tend to have +3 or 4 attack bonuses. Save bonuses, OTOH, go down to -1 or lower, and may never increase. A 20 AC isn't breaking BA (a 25 would be). A 20 save DC is. What's more, the high AC breaking BA is pretty theoretical, it'll impact a 1st level character in a sandbox game who's chomped on more than he can bite off, let alone chew, while the high save DC breaking BA just requires a save vs a non-proficient, lower-priority-for-that-character stat, of which there may be up to 4 out of 6...

 

[FormerlyHemlock]

I don't see how it can be: The exact same thing could have happened in 3.x, in which accuracy was wildly un-bounded, for instance.

I think part of the problem we're having is seeing only the bonus side of the equation as 'bounded,' and perhaps, also in seeing only the upper end of it as bounded. AC and save DCs also need to be bounded, for instance, and bounded relative to plausible bonuses used to hit them. Attack bonuses are typically an attacker's best bonus. They pick the weapon that sues their STR or DEX, whichever is higher and in which they're proficient or use a cantrip/spell which'll generally be off their best stat and always include proficiency. Even really weak fractional-CR critters tend to have +3 or 4 attack bonuses. Save bonuses, OTOH, go down to -1 or lower, and may never increase. A 20 AC isn't breaking BA (a 25 would be). A 20 save DC is. What's more, the high AC breaking BA is pretty theoretical, it'll impact a 1st level character in a sandbox game who's chomped on more than he can bite off, let alone chew, while the high save DC breaking BA just requires a save vs a non-proficient, lower-priority-for-that-character stat, of which there may be up to 4 out of 6...

I must have overlooked the post where you quoted Rodney Thompson too, but this post reveals that you seem to be using a wildly different definition of "bounded accuracy" than he does. AC 25 doesn't "break" bounded accuracy. (Clearly not--both because 5E monsters sometimes have ACs that high, and because you can wind up fighting and even beating AC 25 at first level, before any advancement even occurs.) AC 35 on a monster would probably be a violation of bounded accuracy because it reveals a DM who apparently expects PCs to have lots of plusses to-hit; it would not be a violation of bounded accuracy if the DM was intentionally expressing the monster as being almost impossible to hit, e.g. a diamond-hard creature the size of a pin-head might appropriately have AC 35, and if any PCs manage to hit it that's a bonus due to their sheer awesomeness and not just a difficulty treadmill the DM set up to "challenge" high-level PCs.

Countering Meteor Swarm with Counterspell is bounded accuracy at its purest because of bounded accuracy is all about increasing in HP, damage, and special abilities instead of just pluses to your d20 roll. Counterspell is a special ability. The fact that it could have happened in 3.x doesn't mean anything; it's not like the principles of bounded accuracy were invented out of whole cloth. AD&D had most of the same principles.

What Bounded Accuracy is very definite about is that there shouldn't be any DM-side assumptions about the PCs increasing their offensive or defensive d20 bonuses as they level up. It's not a treadmill. Maybe you like treadmills and want to throw BA out the window--many of the DMs on this thread apparently do--but don't try to justify it in the name of a completely different principle which you name "bounded accuracy" just to be confusing. If you use save DCs of 20+ simply because you're running high-level adventures and want to "challenge" PCs with high bonuses, and then you're distressed because the PCs with low bonuses never make the saves... the problem was created not by 5E but by the fact that you, the adventure designer, are violating the assumptions of bounded accuracy. You've re-invented the treadmill.

You could call it "Unbounded Accuracy 5E" though. If that's what you like, more power to you, and I'll stay off that thread.

 

[CapnZapp]

BA isn't perfectly-implemented (what can be, in the 5e philosophy of standard-rules-as-starting-point?), but the same issue ("saves"* net getting worse as you leveled) has existed for a while: even more dramatically in 3e, in which accuracy was radically unbounded, and to an extent, in 4e, where it was rapidly-but-evenly scaling. It's been a major change to the game's dynamics that's rarely addressed - like the trend of removing limitations from casting with every single edition, or the 5MWD....











* scare quotes necessitated by 4e not technically using 'save' the same way, but having 'non-AC' defenses - darn that non-conforming red-headed step-child of an edition. ;P

Ah, finally a short enough post!

Just wanted to chime in saying that the issue where you get progressively worse at saving has a considerable upside: spells like Hold Person, which are pretty lame at low level, start to rock at high levels!

The problem is ONLY that the game allows for (doesn't care enough to prevent) impossible saves.

Sent from my C6603 using EN World mobile app

 

[Tony Vargas]

I must have overlooked the post where you quoted Rodney Thompson too, but this post reveals that you seem to be using a wildly different definition of "bounded accuracy" than he does.

I don't think so. 5e keeps the swing between best & worst bonuses relatively narrow, enough so that it doesn't overwhelm the d20. OTOH, it scales hp/damage dramatically. I wasn't aware of the specific quote, but it fits what I've said about BA from the beginning. In simple race-to-0-hp combat, it really doesn't matter too much if your opponent has a slightly higher AC than you're calibrated for or your own is a little low (though that's not so hard to fix), it's hp/damage (and numbers) that really make the difference. A low level monsters isn't a non-threat because it can't hit you, but because, even if it does, it doesn't do much damage.

The relative lack of scaling of attack bonuses and AC doesn't render level meaningless, because attacks inflict hp damage, and both damage potential and max hps scale with level.

Countering Meteor Swarm with Counterspell is bounded accuracy at its purest because of bounded accuracy is all about increasing in HP, damage, and special abilities instead of just pluses to your d20 roll.

Meteor Swarm does plenty of damage, so, if you don't bring counterspell into it, it's a fine example. Almost any character might get lucky or unlucky and make (very unlikely if you're talking dumped DEX & no proficiency) or fail his save vs a Meteor Swarm, at least a PC's with only 19 DC, which is BA (barely) working, the d20 is not overwhelmed, but the 1st level character who rolls a 20 on his save is still going to sublimate in the wrath of the spell's 40d6 damage, while the 20th level character who fails his (even if he also fails on a 19), just might be able to take the average 140 damage, a fellow wizard would need a decent CON to avoid being instantly killed by that much damage, while a high-CON fighter or barbarian at full hps might still be standing.

Abysmal saves, like the more theoretical abysmal AC, can work, as long as scaling still comes into it, somewhere - and, in 5e, the one thing that scales for all characters is hps.

If you use save DCs of 20+ simply because you're running high-level adventures and want to "challenge" PCs with high bonuses, and then you're distressed because the PCs with low bonuses never make the saves...

There's no need to always use such high DCs, but it wouldn't be any more in keeping with bounded accuracy to never use them, either.

the problem was created not by 5E but by the fact that you, the adventure designer, are violating the assumptions of bounded accuracy.

That wouldn't actually make it less of an issue, nor give me any less reason to mod it.

Counterspell is a special ability.

It's a spell, one of many. Casters do have the flexibility to pull some kind of solution out of their sleeve for almost any problem, be it narrative or meta-game. That's an entirely different issue...

The fact that it could have happened in 3.x doesn't mean anything; it's not like the principles of bounded accuracy were invented out of whole cloth. AD&D had most of the same principles.

Both 2e AD&D and 3e had 'unbounded' accuracy, specifically in literal 'accuracy,' THAC0/BAB. (3e went particularly far beyond the pale with skill checks, which could have bonuses far in excess of 20, and even over 100.) 1e's matrixes were a little more nuanced, with strings of natural 20s before needing natural 20 + bonus to hit very low negative ACs keeping the d20 relevant longer, but, only on the extreme of needing a 20. The classic game idiom of low-level monsters becoming 'useless' was something 5e was specifically trying to get away from with BA, and prettymuch did, on the AC/attack side. Lower-level monsters can hit the PC in full plate, but if he's high level, they just won't do much damage relative to his massive hps.

What Bounded Accuracy is very definite about is that there shouldn't be any DM-side assumptions about the PCs increasing their offensive or defensive d20 bonuses as they level up.

It would be foolish not to account for increasing offensive bonuses and save DCs, since they're damn near automatic (a player would have to willfully avoid ASIs to his attack stat and choose to use non-proficient weapons to avoid any sort of scaling to attack - one using spells/cantrips literally couldn't avoid scaling, because the proficiency is automatic). Short of such willful sub-optimization, attack bonuses and save DCs are going to scale about as fast as BA allows. Similarly, AC is not that hard for any class to get, even those that get no armor at all usually have some alternate AC boost, even if it uses a spell slot.

It's not a treadmill. Maybe you like treadmills and want to throw BA out the window

The treadmill and BA both addressed the issue of the d20 being overwhelmed. There's a bit of a treadmill in 5e (everyone does get the same proficiency bonus), just with such small numbers it's not as critical you stay on it. The more significant scaling, per the above, is in hps/damage rather than DCs/bonuses.

But that breaks down where only DC/bonuses matter. With skills, that's less of an issue, since experts can usually use skills while amateurs 'help' them, or character can simply gravitate to the skill they're good at. Since saves are not so proactive (indeed, they can be quite flexible/proactive on the side of the attacker forcing the save), it's more of a problem. Having /six/ separate saves exacerbates it.

BA is kinda-sorta-new in that every prior edition has overwhelmed the d20 (or whatever die they were using) with level scaling, and most* have done so with experts vs amateurs (sometimes regardless of level, in the form of 'trained' only or exclusive 'special ability' percentages or the like). 5e's the first to try to avoid the issue in both cases. In doing so, it's made high-level non-experts as incompetent as low-level non-experts, which makes sense with things like tool proficiency and other non-adventuring skills that you're not just going to 'pick up' in the constant struggle for survival that give everyone more hps and higher proficiency bonuses as they level up, but doesn't make much sense for, for instance, AC (why don't PCs get better at dodging by adding proficiency to AC), at least not as obvious sense (because they get bunches of hps as they level, that broadly represents 'dodging' among many other not-getting-killed factors). Since saves are little more, mechanically, than inverted attack rolls, they have a similar issue, it does make sense that you'd get better at resisting magic and avoiding poison and other dangers that are arbitrarily resolved by saves rather than attacks, but that could also all be rolled into hps - iff all failed saves resulted in hp damage or a max-hp or current-hp 'test' like that of the Sleep spell. Since that's not the case, it makes sense for saves to scale - but only a minority of them do, and that's what I don't care for.

I often contrast it with saves getting genuinely better in 1e, but the solutions we occasionally briefly touch upon between attempts to deny the issue or affix blame to the hypothetical DM experiencing it, are closer to treadmill-'better' (that is, at best, keeping pace with rising DCs, and getting 'better' only relative to revisiting past challenges - something that, under BA, can happen quite a bit).

I'd be open to a 'really better' solution, if a balanced one could be found. The equivalent of 1e advancement, scaled down to BA, might, for instance, be accomplished by removing proficiency from save DC calculations, and adding proficiency scaling (+4 over 20 levels), if not full proficiency, to all saves, across the board.






* three guesses.