Guardrails: avoiding unexpected breakage vs preventing exploits

[kyoryu]

Hogwash, people play the modules with the same encounters all the time, with parties that commonly map to things like "2 warriors + mage + priest + 5th wheel". There's no real D&D tactics discussion because nearly all martial builds only really get one effective option and everyone else has the top 3 spells for their class & level and they're used the same way nearly every time.

I'll also add that D&D, as a rule, is a game where your build matters far more than your tactics. Any relevant tactics are either obvious (don't have the mages tank), outshined by build effectiveness, or effectively would only matter if the GM was actually using tactics against you.

Side note: I think D&D would be a better game if it were organized such that:

a. Failure at the encounter level was common
b. GMs were expected to play hard with the resources they had in a given encounter

[Pex]

I can agree D&D is not heavy tactical, but it has some. Not liking that little is a matter of personal taste not game functionality. "Beer & pretzels" gaming is a thing. Certainly for those into heavy tactics with precision, planning, and formations have their fun with it, but other players don't want to have to concentrate that hard. Go with what works and the occasional hiccup to need to do something else is all the fun they need.

[kyoryu]

I can agree D&D is not heavy tactical, but it has some. Not liking that little is a matter of personal taste not game functionality. "Beer & pretzels" gaming is a thing. Certainly for those into heavy tactics with precision, planning, and formations have their fun with it, but other players don't want to have to concentrate that hard. Go with what works and the occasional hiccup to need to do something else is all the fun they need.

True, but (heavily version-dependent statement inbound) I think that the tactics in D&D are either a) incredibly obvious and general or b) tightly bound to a specific build.

I saw more tactical advice around 4e, which upped the tactics more than previous games, than around other editions. (Defining "tactics" as "movement/positioning/etc." vs. "abilities to use in order").

[Xervous]

A big part of D&D’s appearance of shallow tactics is that most maps end up far too close to featureless white planes. You can do the terrain justice, but that tends to expose how melee only is generally a failed concept.

[PhoenixPhyre]

A big part of D&D’s appearance of shallow tactics is that most maps end up far too close to featureless white plains. You can do the terrain justice, but that tends to expose how melee only is generally a failed concept.

I make no claims about being a tactical DM. I'm not. Not at all. And would rather not play a heavily-tactical TTRPG, preferring to leave that to playing things like XCOM with playing pieces I'm not invested in at all in a disposable setting. The war game experience is just not something I care for.

That said, I do try to include terrain, and melee characters tend to be way more useful and important than ranged ones. The fighter and monk in my current game are way ahead on the "important targets taken down" metric compared to the (homebrew, but basically wizard-ish) caster, even including assists with disabling spells. And the druid is basically at zero since she's a support caster by choice. The melee have carried the group, really.

The caster has excelled at taking down groups of smaller foes though. So he does play an important role. Just a different one.

[Xervous]

I make no claims about being a tactical DM. I'm not. Not at all. And would rather not play a heavily-tactical TTRPG, preferring to leave that to playing things like XCOM with playing pieces I'm not invested in at all in a disposable setting. The war game experience is just not something I care for.

That said, I do try to include terrain, and melee characters tend to be way more useful and important than ranged ones. The fighter and monk in my current game are way ahead on the "important targets taken down" metric compared to the (homebrew, but basically wizard-ish) caster, even including assists with disabling spells. And the druid is basically at zero since she's a support caster by choice. The melee have carried the group, really.

My experience whenever I run a moderately competent TTRPG OPFOR is that switch hitters are the typical MVPs. Melee gets to pick off stragglers and generally waits for someone else to reduce or divide the enemies below the critical point where it’s actually safe to join the main fight at full efficiency. A troll street samurai, a 3.5e THF gish, or a WFRP knight are definitely A grade melee kits. The gish can fireball the goblins rather than charging their fortified position, the troll sammy brought axe+minigun because you don’t want to run up to the cybered super cops when they’ve got VTOL gunships hovering outside, and the knight brought a bow. Melee only characters get to sit on their thumbs or collect projectiles/status spam from the firing line.

I should note that my players overwhelmingly choose to pursue hard quests and contracts. Melee-exclusive is the first archetype to crumble when pushing content like this, and they seem implicitly aware of that.

So many games outside modern D&D set a low cost for competent switch hitters. It’s only with STR/DEX split and D&Ds method of stat progression that various STR archetypes get hedged out.

[kyoryu]

I should note that my players overwhelmingly choose to pursue hard quests and contracts. Melee-exclusive is the first archetype to crumble when pushing content like this, and they seem implicitly aware of that.

Melee-exclusive is a terrible archetype, in general, without a lot of abilities to make it viable.

I think that a good fighter archetype should be closer to Batman than anything - canny, resourceful, and able to use a variety of tactics based on what they're encountering.

If you really want an archetype that's that heavily melee-based, they need to have some abilities specifically to counter range advantages - either ways to slough off ranged damage, make better use of cover, extremely high mobility, etc. Or, just, you know, give them other options. Conan throwing swords and knives doesn't make him less barbarian-like to me.

Fighter is often a weird archetype in D&D land (specifically) because it's the one archetype that, more than anything else, is defined by what it doesn't do.

[RandomPeasant]

Melee-only is fine if the melee-only builds can reliably get maneuverability advantages. If your typical melee character is a superhero-style "flying brick", that works fine because the flying brick is also the most mobile character in most fights. If your typical melee character is Sword-o, the regular guy with a sword, you're going to have a bad time the second you add, like, manticores to your game because they will hit Sword-o with tail spines and he will sit there crying because his sword is useless against flying enemies.

As far as the original premise of the thread, I honestly think that the idea of "guardrails" is overestimating how good designers are at balancing games. Most of the broken things in TTRPGs are the result of designers writing an ability and not understanding they have written it with an infinite loop (e.g. Bloodzilla in Shadowrun). In all of the D&D cheese I know across all the various editions, the only thing that springs to mind as "this broke because designers didn't understand how the different pieces of their system interacted" is the d2 Crusader. In software terms, TTRPGs are typically operating without unit tests, and in that environment any sweeping statements you make about the system are going to be coincidentally correct at best.

[kyoryu]

If you really want an archetype that's that heavily melee-based, they need to have some abilities specifically to counter range advantages - either ways to slough off ranged damage, make better use of cover, extremely high mobility, etc. Or, just, you know, give them other options. Conan throwing swords and knives doesn't make him less barbarian-like to me.

Melee-only is fine if the melee-only builds can reliably get maneuverability advantages.

Your tone sounds like it's disagreeing with me, but you're just expanding on one of the things I said.

[KorvinStarmast]

As far as the original premise of the thread, I honestly think that the idea of "guardrails" is overestimating how good designers are at balancing games. Most of the broken things in TTRPGs are the result of designers writing an ability and not understanding they have written it with an infinite loop (e.g. Bloodzilla in Shadowrun). In all of the D&D cheese I know across all the various editions, the only thing that springs to mind as "this broke because designers didn't understand how the different pieces of their system interacted" is the d2 Crusader. In software terms, TTRPGs are typically operating without unit tests, and in that environment any sweeping statements you make about the system are going to be coincidentally correct at best.

Similar to software, though: as the number of features (and in D&D, also spells) increases the number of things that can interact in a non-linear (or other disruptive) fashion increases greatly.
Also, as with software, play testing is required to find some of the problem areas. But you are so right, the depth and degree of play testing most TTRPG dev teams can't afford (as a practical matter).

The above a part of the reason I have for disliking spell bloat. (And n ow I need to look up the D2 Crusader and figure out what you

Put another way, you don't need a spell list in the hundreds.


In Empire of the Petal Throne, the number of spells in the game was not that great (IIRC, a few dozen, I can check later) and spell use was (as compared to WotC D&D) quite limited (eyes were the frequent mitigation to that). Even so, the players were able to have a great deal of fun and full on swords and sorcery + residual SF feel.

[NichG]

I guess my view of guardrails is that you have a few things in the rules where you write explicitly what the intended outcome and dynamics are. Not in absurd detail like 'characters should have a 60-85% chance to hit equal level enemies', but more broadly like 'in any sort of conflict, there should be a monotonic push towards resolution, with abilities and actions serving to guide where that resolution lands, not to prevent it or cause stagnation' or 'anyone no matter how strong should still be threatened by sufficient numbers of weaker individuals'. Then just say, again explicitly, that these intentions are more important than the particular mechanical options and the rules for how they combine. Establish in advance that while the players have the freedom to choose their options and actions, it is also their job to respect those lines, and if something goes over the line it is expected that the specific mechanics will be changed to bring it back into alignment with those lines.

For example, I played with a GM who ran an intentionally gonzo game. Bring your own system, bring your own sources, and he'd integrate them on the fly in a haphazard way and off you go. Homebrew? Even homebrew you wrote? It's fine! Have one character playing by 3.5e D&D rules and another by 4e rules? We'll make it work! Want to throw in a World of Darkness Exalted character? A Nobilis character? Sure, he'd price it out and off you go.

But he had a couple of these guardrails (mixed in with table social contract stuff, so I'll pick out the more 'mechanical' ones):
- One was, explicitly, no infinite combos. Anything that returned a result of 'infinity' or 'arbitrarily high' (e.g. nigh-infinite or NI from charop parlance) just wouldn't work.
- Another was, if you have something cool that makes it hard for others in the group to contribute or participate, everyone else will start getting handouts and opportunities that will catch them up. So you are allowed to use all sorts of system mastery to get yourself things you think are cool, but you're not allowed to use system mastery to justify being cooler than someone else at the table. While it wasn't your responsibility in that case to hold yourself back, no complaining when after you build Pun-Pun someone else finds out that they were secretly the inheritor of the power of an overgod and ends up with 25 free divine rank or something like that.

That's the kind of 'guardrail' I think it makes sense to have even in a system document. Say what the rules are broadly trying to achieve, and say that any interpretation of the rules that would end up causing that to fail just means that the rules are spot-edited, refunds happen as necessary, and you move on.

[gbaji]

Similar to software, though: as the number of features (and in D&D, also spells) increases the number of things that can interact in a non-linear (or other disruptive) fashion increases greatly.
Also, as with software, play testing is required to find some of the problem areas. But you are so right, the depth and degree of play testing most TTRPG dev teams can't afford (as a practical matter).

I think this highlights a point about classes in games and how they *can* operate as guardrails. They work well if there is a relatively small number of classes available and the classes are relatively broad. They serve the purpose of giving new players an "easy way to build a character they want", and ensure that each of these broadly created characters will function "well enough" within the context of the game when played.

Where class based systems get into trouble, IMO, is as the number of classes grows, so as to create more detailed variation between one class selection and another. This appears on first glance to be a natural progression of the game design. Start with a few very broad/basic classes, and then add more variation so that players can customize more exactly what they want. Except what inevitably happens is that these new classes introduce explots. As the number of special abilities designed to differentiate one class from another similar one, grow, the number of potentially exploitable interactions between those abilities grows as well. And that's before considering the fact that it's quite common for the folks writing the expansions which include these new classes just don't spent the time/effort considering how these new abilities they are creating might be used in the first place.


I don't know if there's a single "right/best way" to do this though. I think that if you want "simple and easy to get in and play" then class based systems work great. If you want "greater variation", then I think that skills/ability based games, with a range of well thought out and balanced skills/abilities/spells/whatever, work better. Those latter style games tend to have a higher buy-in cost in terms of learning curve to play, but allow for balanced variation right out of the box. That's not to say that later expansions/revisions which may introduce new things into said games can't also break things anyway, but IME the number of new "things" introduced is often less significant in a more free form classless game system. So those game *can* go off the rails, but I've yet to see a class based system where at some point, the introduction of more and more classes and sub-classes and whatnot don't cause significant problems (and usually require GMs making hard decisions in terms of what is actually allowed in the game).


It's also telling that most expansions for skills/ability based games are more focused on introducing new settings/locations/NPCs/organizations/whatever to the game to be used as a resource, than they are at introducing new core abilities and game mechanics. Well made classless games simply don't need to add a whole lot of that stuff over time, so they tend to focus on setting/adventure stuff rather than "new choices for players to run".

At least, that has been my experience. Again, there are certainly examples that can run counter to this, but as a general trend, this is what I've observed.

[Satinavian]

I don't know. That argument boils down to "If you have many things you can freely combine, unexpected results can happen. If you heavily limit the combinations, you can avoid this".

That is obviously true. But it is also true that when a system restricts combinations enough that the designer can easily foresee every possible character, then so can the GM and the players, making the system restrictive enough that no surprising or unexpected character can ever exist.


That is not guardrails, that is a trade-off. And thus just a design decision about valuing one thing or another.

[kyoryu]

I don't know. That argument boils down to "If you have many things you can freely combine, unexpected results can happen. If you heavily limit the combinations, you can avoid this".

That is obviously true. But it is also true that when a system restricts combinations enough that the designer can easily foresee every possible character, then so can the GM and the players, making the system restrictive enough that no surprising or unexpected character can ever exist.


That is not guardrails, that is a trade-off. And thus just a design decision about valuing one thing or another.

For sure. It's all tradeoffs.

This isn't binary, though. You can have a system where classes exist, and have some breadth in them. That significantly reduces the combinations you have to deal with, while offering a decent-to-good amount of customization.

[gbaji]

I don't know. That argument boils down to "If you have many things you can freely combine, unexpected results can happen. If you heavily limit the combinations, you can avoid this".

It's a bit more nuanced than that. If you start with a large number of things, and fully test their combinations, as part of the game design, then the players are free to combine those things, and avoid exploits. So systems that are designed from the ground up as "buy/choose sets of skills/abilities/spells/whatever and make your character as you wish" tend to be pretty good at avoiding exploits (not all of them though!). Since the game was built to be flexible, the initial testing for the game included that flexibility in the testing, so it's less likely to break based on a varity of combinations of those things. Even as new "things" are added to the game, since the core game already incluides essentially a "template for balacing things", it's less easy to accidentally introduce exploits.

On the flip side, a game system designed with a smallish number of classes, with a limited set of "things", can be more easily balanced and tested by just comparing what each class gets as a set. So there may be no need to test for "what if I take ability A from this class, and use it with spell B from this other one". That literally doesn't come up in testing, and isn't in scope, so it's more likely that there will be problems when you start mixing and matching those "things". Then, as new classes are introduced to the game, the likelihood is that there is less testing of those things. And as the number of expansions with new classes gets added, it becomes increasingly each to accidentally introduce exploits. The game, and the "things" within it, weren't literally designed to be balanced except via combination exclusion. So as new classes are added, the combinations which weren't tested previously (or maybe were, but the new writers don't realize it), could be included and break things.

That is obviously true. But it is also true that when a system restricts combinations enough that the designer can easily foresee every possible character, then so can the GM and the players, making the system restrictive enough that no surprising or unexpected character can ever exist.

Sure. Which, I suppose, depends on how much enjoyment of an adventure based RPG is from the details of the character as written on the sheet, versus the personality imparted by the player when running that character, and the various decisions and encounters and interactions the PC has while in the game. Some of us really enjoyed playing "Dungeon" back in the day, despite there only being exactly 4 classes (and each was functionally identical). I consider a character to be a lot more than just a combination of stats and abilities written on a sheet. And I can absoluitely say from many years of tourney GMing, that I can hand 5 different players the exact same character sheet, and I'll get 5 very very different characters played at the table.

That is not guardrails, that is a trade-off. And thus just a design decision about valuing one thing or another.

Eh. Not necessarily. What I was observing is that if the "things" in the game are designed from day one to interact a specific way, and to be balanced based on a variety of different interactions, then you give the players a template to build a wide variety of characters, and (assuming decent testing was done initially), there should be no major exploits (there will certainly be come combinations that are more optimal than others, but that's to be expected in a more open format game system).

The trade off, as you say, is that those systems have a higher "buy in" cost for the players and GM. It can be harder to figure out what does make a good character, or how best to reflect different sets of "things" in a game, when it's more free form. That can turn off players to the game, while a more restricted class based system can get someone up and running in a game very quickly. Which is "better" kinda depends on the players and the GM.

[Pex]

It's a bit more nuanced than that. If you start with a large number of things, and fully test their combinations, as part of the game design, then the players are free to combine those things, and avoid exploits. So systems that are designed from the ground up as "buy/choose sets of skills/abilities/spells/whatever and make your character as you wish" tend to be pretty good at avoiding exploits (not all of them though!). Since the game was built to be flexible, the initial testing for the game included that flexibility in the testing, so it's less likely to break based on a varity of combinations of those things. Even as new "things" are added to the game, since the core game already incluides essentially a "template for balacing things", it's less easy to accidentally introduce exploits.

On the flip side, a game system designed with a smallish number of classes, with a limited set of "things", can be more easily balanced and tested by just comparing what each class gets as a set. So there may be no need to test for "what if I take ability A from this class, and use it with spell B from this other one". That literally doesn't come up in testing, and isn't in scope, so it's more likely that there will be problems when you start mixing and matching those "things". Then, as new classes are introduced to the game, the likelihood is that there is less testing of those things. And as the number of expansions with new classes gets added, it becomes increasingly each to accidentally introduce exploits. The game, and the "things" within it, weren't literally designed to be balanced except via combination exclusion. So as new classes are added, the combinations which weren't tested previously (or maybe were, but the new writers don't realize it), could be included and break things.

Reminds me of a classic 3E oopsie.

A splat book is published. Divine Metamagic feat is created. Allow a cleric to expend Turn Undead uses to fuel a metamagic feat on a spell instead of increasing spell slot level. Powerful but manageable given the limited number of Turn Undead uses a cleric will have.

Another splat book is published for its own niche. Nightsticks magic item are created. Give a cleric more uses of Turn Undead. A useful item for an undead heavy campaign of which the published splat book facilitates.

Individually Divine Metamagic and Nightsticks are cool things to have. Put the two together and the game falls apart not even counting Persistent Spell.

[Witty Username]

I don't have a strong preference between classless and classes. I do have moods and and opinions on what is fit for purpose.

But I would say two big drawbacks of classes is stagnation and bloat. Advancement in a core function is a system expectation of a class system, and so past a determined range you will have these as consequences of that.
Bloat is the obvious one, HP in 5e, attack bonuses in previous versions of D&D. The numbers grow and hit a breaking point in the fiction or system. A control for this is frontloading, where early levels are higher impact than late levels, but this only mitigates the issue rather than remove it, and can exasperate the second issue.
Stagnation, the sense of advancement losing meaning. 5e has this problem all over, the best spells are 3rd-5th level for the most part, new features are a rarity past 11th level. This creates indifference towards high level play. There are two possible solutions here, non-defined lateral gain like multiclassing, or consistent improvement on the supported level range. Both of which run counter to the advantages of a class system, and tend to accept bloat as inevitable.

Classless systems tend to avoid this with an expectation of lateral growth, and a flat removal of most numeric gains. Even systems like Blades in the Dark draw on this, with class being much more a starting point than a route of progression.

[Xervous]

I don't have a strong preference between classless and classes. I do have moods and and opinions on what is fit for purpose.

But I would say two big drawbacks of classes is stagnation and bloat. Advancement in a core function is a system expectation of a class system, and so past a determined range you will have these as consequences of that.
Bloat is the obvious one, HP in 5e, attack bonuses in previous versions of D&D. The numbers grow and hit a breaking point in the fiction or system. A control for this is frontloading, where early levels are higher impact than late levels, but this only mitigates the issue rather than remove it, and can exasperate the second issue.
Stagnation, the sense of advancement losing meaning. 5e has this problem all over, the best spells are 3rd-5th level for the most part, new features are a rarity past 11th level. This creates indifference towards high level play. There are two possible solutions here, non-defined lateral gain like multiclassing, or consistent improvement on the supported level range. Both of which run counter to the advantages of a class system, and tend to accept bloat as inevitable.

Classless systems tend to avoid this with an expectation of lateral growth, and a flat removal of most numeric gains. Even systems like Blades in the Dark draw on this, with class being much more a starting point than a route of progression.

This appears to be a partial conflation of the classed nature of the system with its chosen dice mechanic. Dicepools like those seen in BitD provide decreasing returns for investments past the point of competence. For D20 a +1 is forever +5%. A hypothetical D&D where you assign level up bonuses a la carte still runs into the fact that +1 is a consistent adjustment to your success rate until you’ve reached the bounds of the D20. Flipping additional coins looking to get at least one heads quickly sees the gains plummet after you hit 3 coins. What value is a 6% success rate gain here when adding a coin to a different pile may be establishing its 50% success rate (up from zero) or pushing a single coin pile from 50% to 75%? In D&D any of those attempts to broaden a character’s competencies run into the D20, +1 remains a 5% whether it’s nudging your underwater basket weaving or your murderous sword arm.

[gbaji]

Stagnation, the sense of advancement losing meaning. 5e has this problem all over, the best spells are 3rd-5th level for the most part, new features are a rarity past 11th level. This creates indifference towards high level play. There are two possible solutions here, non-defined lateral gain like multiclassing, or consistent improvement on the supported level range. Both of which run counter to the advantages of a class system, and tend to accept bloat as inevitable.

To be fair, original AD&D had the same "problem" (if we call it that). I always found that the most enjoyable level ranges to play was in the 5th through 8th levels. Below that, characters feel like they are out of their depth a lot of the time, and "death to minor foes" could happen frequently. Above that, it seemed like the power scale just got off kilter and didn't work well. Within that range though, the scope and scale of the characters relative to what most considered an "heroic" environment, just worked well. This is also the reason why most tourney scenarios were set in this range back in the day as well. It allowed for PC parties that had enough power and flexibiility to be capabile of handling a wide assortment of challenges, but hadn't yet hit the more "problematic" power ranges yet.

Classless systems tend to avoid this with an expectation of lateral growth, and a flat removal of most numeric gains. Even systems like Blades in the Dark draw on this, with class being much more a starting point than a route of progression.

Eh... The two are strongly interconnected, but I'd argue it's the Level based system component that is the real problem here, and not the very existence of classes. But yeah, most games which have classes also have levels, so that may be a moot point.

But, IMO, what drives the linear rather than lateral growth of PCs is the game mechanics associated with the levels of characters and specifically creating level equivalents for opponents for them to conflict with. When the level of opponents is scaled to the level of the party, and that level determines how difficult they are to affect in some way (to-hit rolls, opposed skills, spell resistance, whatever), it drives the PCs to focus on maximizing the equivalent value of their respective skills/abilities (so as to be effective against the opposing values of "level appropriate" opponents).

Which really encourages a narrow focus on the advancement itself. So even if you introduce skill points into a level based game, there's going to be a tendency towards focusing on maximizing a small number of skills instead of speading those points around. This is precisely what happened with all editions of D&D that have included skills (which is basically every one past 1st edition). You get X skill points per level, and you really have to keep spending most of them on the same set of skills you've been focusing on since level 1, or you will "fall behind" in the effectiveness of those skills against any opponents you will encounter.

This appears to be a partial conflation of the classed nature of the system with its chosen dice mechanic. Dicepools like those seen in BitD provide decreasing returns for investments past the point of competence. For D20 a +1 is forever +5%. A hypothetical D&D where you assign level up bonuses a la carte still runs into the fact that +1 is a consistent adjustment to your success rate until you’ve reached the bounds of the D20. Flipping additional coins looking to get at least one heads quickly sees the gains plummet after you hit 3 coins. What value is a 6% success rate gain here when adding a coin to a different pile may be establishing its 50% success rate (up from zero) or pushing a single coin pile from 50% to 75%? In D&D any of those attempts to broaden a character’s competencies run into the D20, +1 remains a 5% whether it’s nudging your underwater basket weaving or your murderous sword arm.

That's not 100% correct (but close enough, I suppose). In theory, as long as you keep the "scaling windoow" of difficulty for a D20 system within a nominal 20 point range of relative difficulty, then a +1 is a +5%, and the scaling works (yeah, yeah, relative increase odds are different... blah blah). If you don't, then you do get a minimized return on investment. If the difficulty window never changes, then once you hit a "I succeed on any roll other than a one", there's no value to increasing the points in the skill. The problem is if we are considering some kind of opposed roll dynamic, if we don't scale the enemies up, then the PCs will simply buy skill points to where they will "auto win" every time (but will stop at that point, which I suppose allows for lateral skill growth at that point). But if we do scale enemies up, then the need to keep increasing skill points is ever present for the PCs, and will almost require them to keep spending those points on linear growth instead.

It would be interesting to see if one could insert some sort of actual increasing cost versus return math into a game like D&D. But to do this, you would have to also apply a similar diminishing increase rate for NPC opponents as they gain levels. I'm not sure the sheer volume of hacking the system to bits to make this work would be worth the effort though, since those values are embedded in like everything in the game system.

I'm not sure if this fits into the topic in terms of "preventing exploits" though, unless we consider the problem of PCs "auto-winning" an exploit. It's certainly a problem though. And I think the real trick, if we're looking at relative skill levels, is how to scale these upwards, while both retaining a reasonable boundary at any given experience "level", but also allowing for reasonable and matching scaling of opponents at the same relative "level" (I'm putting "level" in quotes, because this need not be defined by actual game levels, but just relative assessments of "how powerful/skilled is someone"). And yeah, having actual level based advancement makes this far far more difficult IMO.

[RandomPeasant]

Similar to software, though: as the number of features (and in D&D, also spells) increases the number of things that can interact in a non-linear (or other disruptive) fashion increases greatly.

I think people significantly overestimate this as a source of brokenness. Most broken mechanics I have encountered aren't a result of some weird interaction between four different things from three different books. They're a result of writing an ability that does something broken. planar binding doesn't break 3e D&D because there's a weird outsider in the MMV that throws off the CR/HD curve that the spell relied on for balance. It breaks 3e D&D because the basic thing that it does is "allow you to recruit as many minions as you want that are individually as strong as any member of your party".

Maybe this is more of a problem in classless systems, but frankly I see that as a problem with classless systems. If you understand what your system is trying to do, you can write classes that do those things, and if you write those classes in a halfway-competent way, you can avoid most of the "it turns out these unrelated abilities form Voltron" problems. Look at a game like MTG. They have tens of thousands of cards, and even the largest formats have small banlists, where many of the entries are a result of overtuning rather than obscure interactions.

The above a part of the reason I have for disliking spell bloat.

Honestly I don't think this is really the problem with spell bloat. If you look at 3e, there's a huge amount of spell bloat, but the number of places where two balance spells interact to do a broken thing is basically zero. The problem with spell bloat is that, when it is not accompanied by an increase in the number of classes, it results in casters rapidly outstripping everyone else in terms of how many options they get. And even that is only a problem for certain ways casters can work (that is: the Sorcerer's "you know X spells" is fine, the Cleric's "you know all the spells at each level" is broken, and the Wizard's "you can learn new spells by finding items" works if you can impose limits).

Put another way, you don't need a spell list in the hundreds.

Well, that depends on what you want from spells. I can certainly think of hundreds of spells I would like to have in the game, but they don't all need to be accessible to the same character at the same time. That said, I think many people over-correct and ask for spell lists to be cut too short.

I think this highlights a point about classes in games and how they *can* operate as guardrails. They work well if there is a relatively small number of classes available and the classes are relatively broad. They serve the purpose of giving new players an "easy way to build a character they want", and ensure that each of these broadly created characters will function "well enough" within the context of the game when played.

I don't think you even need the list of classes to be small (you do want broad classes, but that's just so that you don't have a game that makes it impossible to implement concepts before the right splatbooks are released -- 4e had a huge problem that resulted from making Barbarian and Bard splat content). You just need classes to be modular elements that can be tested in isolation. If you can look at a Scout or a Fire Mage or a Knight or an Oracle and tell if it is balanced without cross-checking against every other class and every non-class option, you can crank out classes. And this is quite doable with a combination of well-defined balance targets, reasonable tolerance for imbalance, and willingness to errata major outliers.

Except what inevitably happens is that these new classes introduce explots.

Could you speak to the examples of this you see? Because I really can't think of any cases where this happens at all.

And that's before considering the fact that it's quite common for the folks writing the expansions which include these new classes just don't spent the time/effort considering how these new abilities they are creating might be used in the first place.

Well, sure, but you can't design a set of guardrails that's robust against people not using them.

Reminds me of a classic 3E oopsie.

A splat book is published. Divine Metamagic feat is created. Allow a cleric to expend Turn Undead uses to fuel a metamagic feat on a spell instead of increasing spell slot level. Powerful but manageable given the limited number of Turn Undead uses a cleric will have.

Another splat book is published for its own niche. Nightsticks magic item are created. Give a cleric more uses of Turn Undead. A useful item for an undead heavy campaign of which the published splat book facilitates.

Individually Divine Metamagic and Nightsticks are cool things to have. Put the two together and the game falls apart not even counting Persistent Spell.

I don't think this is a good example, because I don't really think DMM + Nightsticks is the issue. It's the easiest way to abuse it, but the real problem is DMM + Persistent. Which is just "if you are going to write an effect that reduces the cost of metamagic, you need to make sure it works with all the metamagic your system has", and I don't really think missing something that obvious is forgivable. Like I said, the d2 Crusader is the best example of this, combining a "reroll 1s" ability from one book with a "roll an extra die if your first die rolled its max" ability from an unrelated book to do literally infinite damage. And the fact that there's one really clear example of this in the edition that went the hardest for a content explosion suggests to me that this is not really the problem with content explosions.

[Pex]

I don't think this is a good example, because I don't really think DMM + Nightsticks is the issue. It's the easiest way to abuse it, but the real problem is DMM + Persistent. Which is just "if you are going to write an effect that reduces the cost of metamagic, you need to make sure it works with all the metamagic your system has", and I don't really think missing something that obvious is forgivable. Like I said, the d2 Crusader is the best example of this, combining a "reroll 1s" ability from one book with a "roll an extra die if your first die rolled its max" ability from an unrelated book to do literally infinite damage. And the fact that there's one really clear example of this in the edition that went the hardest for a content explosion suggests to me that this is not really the problem with content explosions.

It's not just Persistent spell. You can Quicken your highest level spells. With Silent Spell cast Silence and go to town while other spellcasters are impotent. Maximize makes cleric damage spells more potent and does wonders for healing.

[Satinavian]

You can Quicken your highest level spells. With Silent Spell cast Silence and go to town while other spellcasters are impotent. Maximize makes cleric damage spells more potent and does wonders for healing.

I would say that was kinda the intended effect.

I mean seriously, what else do you think the writers intended divine metamagic to be used ?

[RandomPeasant]

It's not just Persistent spell. You can Quicken your highest level spells. With Silent Spell cast Silence and go to town while other spellcasters are impotent. Maximize makes cleric damage spells more potent and does wonders for healing.

Quicken is the only one that comes particularly close to Persistent in brokenness. "I can do my magic in a silence" is really not exactly big cheese when has a 20ft radius, psionics exists, and most monsters have SLAs anyway. Honestly I struggle to really imagine a situation where "get in close with a caster and then cast spells on them while we are both in a silence" is better than the alternatives of "cast spells on them from far away" and "get in close with them and beat them to death with a stick while they can't cast because of silence". I really question the idea that Maximize "does wonders for healing" when healing spells famously have the caster level-based part of their scaling come in a flat number rather than more dice.

[Pex]

I would say that was kinda the intended effect.

I mean seriously, what else do you think the writers intended divine metamagic to be used ?

The issue is combining Divine Metamagic with Nightsticks giving the cleric a lot more Turn Undead uses than common. The cleric is spamming metamagicked spells which was a problem for many people in 3E at the time.

Quicken is the only one that comes particularly close to Persistent in brokenness. "I can do my magic in a silence" is really not exactly big cheese when has a 20ft radius, psionics exists, and most monsters have SLAs anyway. Honestly I struggle to really imagine a situation where "get in close with a caster and then cast spells on them while we are both in a silence" is better than the alternatives of "cast spells on them from far away" and "get in close with them and beat them to death with a stick while they can't cast because of silence". I really question the idea that Maximize "does wonders for healing" when healing spells famously have the caster level-based part of their scaling come in a flat number rather than more dice.

If you don't see value in preventing a spellcaster from casting while you still can, that's your issue. A 1st level spell for 13 healing is not nothing. 2nd level spells are 26 healing a pop. 3rd level is 39. It can make healing in combat worth it, and even if you still won't using one spell to heal out of combat when you would have had to use two or three makes a difference.

[Telok]

We played with the nightstick dmm thing. It wasn't a problem if you kept it to a single nightstick. The only issues were when you started threating them like ultra flexible metamagic rods. That's frankly what they were if you went in on dmm + stacks of them, flexible & underpriced metamagic rods. But just using them to run stuff like a few uses of extend and reach metamagics was no problem at all.

[Witty Username]

This appears to be a partial conflation of the classed nature of the system with its chosen dice mechanic. Dicepools like those seen in BitD provide decreasing returns for investments past the point of competence. For D20 a +1 is forever +5%. A hypothetical D&D where you assign level up bonuses a la carte still runs into the fact that +1 is a consistent adjustment to your success rate until you’ve reached the bounds of the D20. Flipping additional coins looking to get at least one heads quickly sees the gains plummet after you hit 3 coins. What value is a 6% success rate gain here when adding a coin to a different pile may be establishing its 50% success rate (up from zero) or pushing a single coin pile from 50% to 75%? In D&D any of those attempts to broaden a character’s competencies run into the D20, +1 remains a 5% whether it’s nudging your underwater basket weaving or your murderous sword arm.

I can see the reading but this misses the forest for a tree.

Say for the sake of argument, we applied 3.5 D&D attack progression (an extreme example to illustrate) to the dice mechanic to Blades in the dark.
By 12th level you would be rolling 12 dice for any attack, That is about 80-90% of a perfect success, and a little lower for a critical success. that is the bloat in effect.
then stagnation, as you correctly Identify, the closed nature does cap the limit of growth, but that means after a certain point the growth has lost meaning.

So we could slow that growth, say 1 dice every 5 levels, to approximate Blades in the Dark's cap of 4. so 1st. 5th. 10th, 15th and 20th. Now you are not gaining any benefit on this line for most of your leveling path. In other words we have warded off bloat at the cost of stagnation.

The specific resolution mechanics will change the speed, but not the trajectory inherent to set progression.

[KorvinStarmast]

Just wondering if the OP has found what they are looking for.

[Slipjig]

A big part of D&D’s appearance of shallow tactics is that most maps end up far too close to featureless white planes. You can do the terrain justice, but that tends to expose how melee only is generally a failed concept.

I think part of the problem there is that most game systems don't interact with terrain in any meaningful way. Most spells ignore cover, even when cover should logically provide some protection. And the most basic defense against archers, hiding behind something, is completely negated when 99% of ranged characters have Sharpshooter. There's also no penalty to using weapons clearly unsuited to the environment, like using a polearm in a narrow hallway.

The only people who really interact with terrain are melee martials, and even then it's almost exclusively in terms of "What terrain is preventing me from getting to my opponent?", with terrain being irrelevant again once they close the distance.

[gbaji]

I think part of the problem there is that most game systems don't interact with terrain in any meaningful way. Most spells ignore cover, even when cover should logically provide some protection. And the most basic defense against archers, hiding behind something, is completely negated when 99% of ranged characters have Sharpshooter. There's also no penalty to using weapons clearly unsuited to the environment, like using a polearm in a narrow hallway.

That's another unfortunate side effect of class based games though. What they gain in "quick and easy definitions for characters" they lose in "flexible encounters". The archer point you made is relevant here. No one uses bows unless they are some kind of archery based build, or creature, or whatever. In fact, we can argue that all characters (PC or NPC) are going to tend to focus on and use whatever skills/weapons/spells that they are "best at", pretty much all the time.

So yeah. If someone is bothering to shoot at you with a bow, they probably have bow feats, so cover beomes irrelevant. If someone is using pretty much any weapon or spell or ability it pretty much means that "this is something my class/creature-type uses, so I'm super good at it" (well, unless we're playing at very low levels).

The only people who really interact with terrain are melee martials, and even then it's almost exclusively in terms of "What terrain is preventing me from getting to my opponent?", with terrain being irrelevant again once they close the distance.

Eh. You *can* introduce terrain effects into any game (yes, even D&D). It's just tricky though, and fraught with potential for creating restrictions for some classes, but not others. Yeah. It may make complete sense for the polearm focused melee fighter to be at a disadvantage when fighting in a narrow hallway, but... how much of a disadvantage? And is that likely to make using their "best weapon" less worthwhile than using something else? If I've stacked up feats with using polearms, you'd have to apply a boatload of negatives to make me switch to a different weapon. Which, yeah... is part of the problem.

And if you do? I'm going to feel like you are going after my build. If the effects/damage/whatever of polarm use was balanced in the game system as a whole with the "can't be used in some situations" restriction, then it'll work. But if me using a polarm is already balanced to "other guy dual wielding short weapons", and we both get simlar/balanced damage/hits/effects as a result, then adding in an additional "mine wont work in narrow hallways but his do", becomes a balance problem that has to be sorted out.

I do agree with you though that, along the way, the designers for D&D basically just threw up their hands when it comes to terrain. Which maybe makes for a less realistic game experience, but at some point, you just have to accept the game for what it is. The problem with actually balancing things with terrain in mind, is that if the GM/players don't use it, then some builds become more powerful as a result. I think the designers just went with "we're going to balance against flat featurelness plains" and were done with it.


The way you solve this problem is to encourage broad rather than deep character builds. That way, you can introduce modest environmental effects (including, but not limited to terrain), and players will be respond to those by making action choice differences with their characters. The decision to switch to different weapon combinations based on conditions works if the characters have multiple different, somewhat similarly effective, configurations available to them. Most game systems go in the other direction though. So it's not surprising that terrain becomes less relevant in them as well.

It's just interesting to me, because I play mostly in RuneQuest, where there is literally zero "balance" between "big weapons" and "small weapons". On paper a greatsword or a polearm is *always* a better weapon. Everything else being the same, they have longer reach, and do more damage. The balance point in the game isn't with some kind of "you can do other things with shorter weapons that balance those out", but literally "you can't always use the big weapons". Sometimes, it's due to terrain. Sometimes it's a matter of "how exactly are you carrying those weapons on you in town without everyone noticing?". All characters who use larger weapons, also develop skills and carry around shorter weapons, precisely for cases when they can't use the big ones. And since there are no classes, there's no particular reason for one character to choose any one weapon over another. Everyone can choose to use any weapon they want. They just have to spend time to become good at it. It's not uncommon for experienced characters to have a half dozen different weapons combinations depending on what they are fighting. Maybe I use greatsword when I need to do lots of damage. Or broadsword and shield sometimes. Or maybe dual wielding soemtimes. Or I have a 1h weapon in one hand, and a thowing weapon in the other. Lots of different combinations, and lots of different reasons to use different ones.

So I know it can be done. It just requires a game system that works well with (and encourages) flexible character builds. Most class based systems just don't though.

[Satinavian]

So I know it can be done. It just requires a game system that works well with (and encourages) flexible character builds. Most class based systems just don't though.

I am not sure i really want this.

Most inspiration for whatever fighting character has their primary weapon. And then some small sidearm for emergencies which which they gor significantly less training on top of it being inferior and where the whole strategy revolves around avoiding situations where they ever need to use it.
Sure, it is only most, not all. There always were some special units or individuals trained on a huge viriety of weapons. But this is not the norm. Weapons are expensive, heavy and unwieldy so you would never take a whole set to war anyway and training takes time.

That does not mean that weapons should not be of situational usefulness. But that should result in their weilders being of situational usefelness, not that your archer batallion grabs spears on a foggy day and is as useful as a spearmen unit.


I would prefer different maneuvers, stances and tactics that can be done with the same weapon.