Are feats necessary?

[unbeliever536]

Really though, are they? On a design level?

If a system has good, modular classes with plenty of options (lol) and allows open multiclassing, what do feats add?

Large numbers of feats seem to be (1) worthless, (2) optional class features, (3) extremely limitted forms of multiclassing, or (4) some combination of the above. I feel like there's nothing feats do that couldn't be done better by something else in a class based system. More and more, I look at the entire feat subsystem as a really bad effort to respond to "advantage" subsystems in point-based levelless rpgs. If I were writing a d20 based system, I would dump feats entirely, absorbing them into other parts of the system designed to fit the level-based paradigm.

[Alcino]

I mostly agree with you: many D&D feats, no matter the edition, are either worthless or boring or just patch up a rule weakness. Feat bloat is a real problem, and also a distraction.

However, I believe that there are character traits that would not integrate well into ability scores, skills or class features. Just plain running faster, for example, or being quite resilient to cold climates, or having a crazy-good memory.

[Vadskye]

Conceptually, you want there two be two main things that determine your character: who you are and what you do. Ability scores, race and class determine who you are. Skills and feats determine what you do.

If you remove feats, you have no ability to represent "this is what I choose to be better at" other than modular class features like rogue talents. Many abilities and talents should not be class-dependent, so this is not a suitable solution unless you strip most aspects of player choice out of the game.

[Cespenar]

Conceptually, you want there two be two main things that determine your character: who you are and what you do. Ability scores, race and class determine who you are. Skills and feats determine what you do.

Classes are what you do. I thought this was obvious.

[WhiteLycan]

Without feats, a level 10 rogue is exactly the same as every single other level 10 rogue. Or if you multiclass, every single 5 rogue 5 fighter is the same as every other one. With feats, you can specialize. Without them, you can't.

[Vadskye]

Classes are what you do. I thought this was obvious.
In fluff, yes. From a game design perspective, no. As WhiteLycan said, (noncaster) classes all provide essentially the exact same features. That means they determine who you are, not what you choose to be good at.

[Zaydos]

Feats are useful in that they allow you to pick up abilities which are not tied to any one class. In a perfectly balanced class system you could do them as class features (like rogue talents) and have a list of universal options, but this would be in effect making all classes have a bunch of class features which are just bonus feat.

Feats are useful alongside features like rogue talents as they allow you to pick up little abilities not directly attached to your class.

Looking at the 4 categories of feats you listed. 1) Where does say Power Attack fall? It isn't really tied to any one class, but also isn't worthless. 2) Feats in group 2 and 3 are still useful as it gives you the customization ability to either go "I'm going to diversify a bit" or "I want to be the best darn rogue there is" and while you could build this into a class system you'd effectively just be giving bonus feats.

And take warblade, binder, and totemist. All three have a modular system of abilities that gives them plenty of options. All three also benefit from generalist feats (Power Attack), feats to specialize in what they do (incarnum feats, improved binding, adaptive style isn't unknown on a warblade), as well as feats to pick up abilities from outside their class's focus (shape soul meld, martial study). I choose these because they actually have abilities that are functionally different (although all possibly filling the melee role, Incarnate could have been used instead of Totemist, or sorcerer instead of any of them).

So ultimately while no they are not necessary, 2e and earlier did just fine without them, they are however beneficial. In making something with modular abilities you must either have spreading out grant you abilities appropriate to your level regardless of your level, and modular enough abilities that you can go specialist in a specific focus of your class, specialist in a general focus of your class, or generalist abilities that let you function outside of your class's specialty. There are drawbacks (especially in the power balance of existing feats) but the benefits of feats outweigh them.

[Cespenar]

In fluff, yes. From a game design perspective, no. As WhiteLycan said, (noncaster) classes all provide essentially the exact same features. That means they determine who you are, not what you choose to be good at.

Meh. The doing/being dichotomy looks inaccurate to me. Every part of a character, after all, provides mechanical bonuses. So you can't really separate one from the other. Unless you're doing fluff, but you refused that notion as well.

Anyway, never mind. Nitpicking and all.

[unbeliever536]

"Are feats necessary" probably wasn't the right way to ask the question. More detail below, but what I'm really trying to get at is: is it possible to do what feats try to do in a way that more closely follows the level-based paradigm of D&D, or must there be some concession to systems that don't follow that paradigm for the sake of character breadth? More on this sort of thing in my specific responses to you guys, below.

Alcino, Vadskye:


I mostly agree with you: many D&D feats, no matter the edition, are either worthless or boring or just patch up a rule weakness. Feat bloat is a real problem, and also a distraction.

However, I believe that there are character traits that would not integrate well into ability scores, skills or class features. Just plain running faster, for example, or being quite resilient to cold climates, or having a crazy-good memory.


Conceptually, you want there two be two main things that determine your character: who you are and what you do. Ability scores, race and class determine who you are. Skills and feats determine what you do.

If you remove feats, you have no ability to represent "this is what I choose to be better at" other than modular class features like rogue talents. Many abilities and talents should not be class-dependent, so this is not a suitable solution unless you strip most aspects of player choice out of the game.

To both of you (and everyone else, but specifically addressing this issue): What if the former functionality of feats (that is, specific options not strictly associated with a particular class) were rolled into a different system, such as skills. You can see some of this in what some people (brewers and late designers alike) do with skill tricks in CS, or with skills generally in Tarkisflux's Tome of Prowess (http://www.dnd-wiki.org/wiki/Tome_of_Prowess_%283.5e_Sourcebook%29). Why shouldn't, say, the excellent athlete, who has invested heavily in the Climb, Jump, and Swim skills, naturally gain the kind of fitness and coordination necessary to execute the manuevers represented by the various running feats (and similar abilities related to his athleticism)? Basically, can we absorb the best parts of the feat system into a more streamlined system for outside-class advancement that better fits the level-based paradigm of D&D?


Lycan:


Without feats, a level 10 rogue is exactly the same as every single other level 10 rogue. Or if you multiclass, every single 5 rogue 5 fighter is the same as every other one. With feats, you can specialize. Without them, you can't.

In 3.5, yeah. I'm asking this on a more fundamental level. My claim, basically, is this: D&D claims to be a level based system, even more fundamentally than a class based system. Unlike many other RPGs, characters advance in discrete jumps, and their abilities all get better at that time. Sometimes, this is represented by a strict numerical buff: increased skill ranks, increased saves, increased bab, etc. Sometimes it's represented by a new or improved ability related to the character's class, but always class advancement represents a character becoming better at doing the thing that their class does.

There's much more space for vertical advancement in a level based system, but somewhat less space for horizontal advancement (at least in theory). A character can choose to take a level in a different class instead of that character's prior class, or allocate their skillpoints slightly differently, but otherwise is pretty much locked. This is the default assumption for most parts of D&D. The sole exception is the feat subsystem. Feats default to horizontal advancement (though through the use of trees they can become vertical); once a character meets the prerequisites, they can take the feat any time they have an advancement opportunity.

I claim that if the level based parts of the system (skills, basically) are robust enough (they aren't in 3.5, pretend for a moment they are), we don't need the feat system to represent the kind of "dabbling" that it tries to. It can be covered entirely by "skill dips" or a similar idea.



Zaydos:


Feats are useful in that they allow you to pick up abilities which are not tied to any one class. In a perfectly balanced class system you could do them as class features (like rogue talents) and have a list of universal options, but this would be in effect making all classes have a bunch of class features which are just bonus feat.

Feats are useful alongside features like rogue talents as they allow you to pick up little abilities not directly attached to your class.

Looking at the 4 categories of feats you listed. 1) Where does say Power Attack fall? It isn't really tied to any one class, but also isn't worthless. 2) Feats in group 2 and 3 are still useful as it gives you the customization ability to either go "I'm going to diversify a bit" or "I want to be the best darn rogue there is" and while you could build this into a class system you'd effectively just be giving bonus feats.

And take warblade, binder, and totemist. All three have a modular system of abilities that gives them plenty of options. All three also benefit from generalist feats (Power Attack), feats to specialize in what they do (incarnum feats, improved binding, adaptive style isn't unknown on a warblade), as well as feats to pick up abilities from outside their class's focus (shape soul meld, martial study). I choose these because they actually have abilities that are functionally different (although all possibly filling the melee role, Incarnate could have been used instead of Totemist, or sorcerer instead of any of them).

So ultimately while no they are not necessary, 2e and earlier did just fine without them, they are however beneficial. In making something with modular abilities you must either have spreading out grant you abilities appropriate to your level regardless of your level, and modular enough abilities that you can go specialist in a specific focus of your class, specialist in a general focus of your class, or generalist abilities that let you function outside of your class's specialty. There are drawbacks (especially in the power balance of existing feats) but the benefits of feats outweigh them.

Power Attack is a wierd one. Given that it's a near requirement for most melee builds that don't take advantage of their own subsystem, and doesn't help anyone else, I'd say it best fits into my category (2) above: an optional feature for a number of melee-focused classes, especially Barbarian and Fighter. There are a number of more obvious instances of category (2), especially in splats, but it works. Power Attack can also be viewed as a very rare exception to category (1).

I'm not experienced with the Binder, so I can't really comment on it. The Totemist and Warblade, however, provide great examples of category (3) feats. Shape Soulmeld and Martial Study, respectively, allow one to, in a very limitted fashion, gain access to the subsystems used by these classes. This is only necessary, though, because of one problem in D&D: subsystems mix horribly. The incarnum and initiator subsystems are actually some of the best when it comes to mixing, in that the player isn't horribly penalized for leaving their class. Incarnum shapers still have essentia caps increase and initiator level still goes up. However, none of the abilities they have improve in a material way. There is no way for a Totemist, for example, to only sacrifice part of his advancement (say, not getting more soulmelds, but still getting increased essentia) for a start at Warblade advancement, unless he uses a feat to pretend to be a Warblade pretending to be a Totemist for a level, which is just wierd. If taking a level in a new class allowed characters to split the level between their two classes, such feats would not be necessary.

It would also mean that the prestige class archetype "you can do both of these things" would be completely unnecessary, which would be a benefit in and of itself. Some of the later ones are ok (usually the ones that let you use one subsystem to power another), but Mystic Theurge is literally just "you are [arcane caster] and [divine caster], but you suck a little less than you otherwise would". Cerebremancer is the same. Why not have doubling up on a class be a semi-reasonable proposition, as opposed to dip-chaining into a prestige class?


And on the doing/being thing: if two classes give identical class features, isn't that bad design? The whole point of choosing a class in a class-based system is that you want an overarching theme for your abilities ("what you're good at"), not just your fluff ("who you are").

[SiuiS]

Yes, feats are required. They've been handled badly, but they are a fantastic selling point and possibly the sine most useful innovation from third edition, which (supposedly) clearly delineated skills and feats out of no weapon proficiency slots.

Adventurer Conquerer king does probably the best job I've seen of utilizing these. I also feel a lot of "feats" would be better off as skill tricks, and skill tricks should be cheaper and less limited.

[NichG]

Feats are one way to do something for which there are quite a few different approaches. I would say that the main problem with feats is this:

- Feats try to be an orthogonal direction of customization, but often there is a very strong connection between the rest of the character's build and what feats they (should/must) choose.

An example of this is PrC prerequisites - by creating a certain set of prereq feats for a PrC, it forces specific choices at specific points if you want to enter that PrC as early as possible. Which means that feats are not actually serving their purpose (in that case) as an extra form of customization.

The second common example is no-brainer feats, where basically this is really core to being functional in certain ways. The Two-Weapon-Fighter feats and actually being a dual-wielder, for example - because they're a given, you can't really consider it customization.

The third common example is feats that are very much 'augments for this/these particular classes'. A Wizard has little use for Power Attack. A Fighter can't get anything from Metamagic feats normally. Someone without a source of Turn Attempts can't really make use of a feat to convert them into other stuff. So basically, feats like this end up being more like ACFs than actual independent directions of customization from your class.

Now, what do feats do well?

- Even for feats that are tied to other aspects of the character, choosing between those feats lets the player make a statement about which of those aspects is most important to them. If you take Metamagics it means you want your spells to have more oomph or versatility. If you instead take racial feats, it means you're focusing a little less on your class and more on your race. So in broad strokes at least, feats do provide a way to shift emphasis.

- Feats are often useful ways to future-proof certain packets of mechanics. If you make a Metamagic feat, that will still work if someone introduces a new kind of caster in homebrew/splats later on. If instead metamagics were tied to the Wizard class, then the 'Symbologist' homebrew would have to copy all those mechanics if it wanted to benefit from them.

So, given these problems and benefits, what are some alternatives?

1. Embrace what seems to be de-facto true, and focus more on choices that you get as part of advancing in a class rather than supposed-to-be-separate-but-not-really choices. In this model, each class would have some set of things you could choose from as you leveled. This would work well in a system with a fairly small number of classes - maybe five or six maximum. Its as if each class got its own unique feats to choose from. It would also be easy for other things about your character to allow you to take alternate options as you leveled, e.g. 'you can invest in the Dwarf power this level or the Wizard power this level, your choice'.

Pros and Cons: You still get customization, but things are tied much more strongly to the other choices. This makes each class a lot deeper and more nuanced, but it starts to feel like 6 separate games if you aren't careful.

2. Paired advancement. This would sort of be like gestalting, except you'd gestalt between two mechanically very different aspects of your character. So for example you could level in 'training' on one side and 'persona' on the other side, or something like that.

Pros and Cons: Kind of a weird system, but combinatorically allows a reasonable degree of customization. Still very linear though.

3. Everything is learnable. In this sort of model, rather than having specific character build resources you get as you level, you have two tracks of advancement - one tied to some overarcing schedule (this would be the 'levels of classes'), the other tied to some per-game and fairly fungible resource like XP, or even just 'downtime'. As you expend this, you can pick up, swap out, etc, etc all sorts of different microabilities. The same way the Wizard can learn a spell, you could learn a maneuver. Or a specialization in a certain kind of attack. Or a new way to wield a certain kind of weapon. This would be something like if feats were spells and you could just build up a library of them with training, but only deploy so many at a time.

Pros and Cons: I kind of like this actually. Cons are, lots of book-keeping and the possibility of everyone ending up with the same 'loadout' for a given scenario. Also what characters can do changes pretty fast. Pros: Versatile, lots of little packets of mechanics to play with, very little cost to making a wrong choice, everyone can end up with 'signature abilities'.

Certainly there are more possibilities than this...

[unbeliever536]

2. Paired advancement. This would sort of be like gestalting, except you'd gestalt between two mechanically very different aspects of your character. So for example you could level in 'training' on one side and 'persona' on the other side, or something like that.

Pros and Cons: Kind of a weird system, but combinatorically allows a reasonable degree of customization. Still very linear though.



I'm idlly working on a system that uses a variation on this alternative, and that's actually the main reason I started this thread. My idea is that each character would, essentially, be a gestalt of a bunch of things:

a) Race: perhaps 1/4 of a well designed D&D 3.5 class or thereabouts.
b) Class: perhaps 1/2-3/4 of a well designed D&D 3.5 class, also determines your access to
c) Proficiencies: Roughly equivalent to a skill + several feats/skill tricks, but covering much more ground as a whole. Each character advances something like 6-8 of these at once, and they're the main place where single-class characters get to differentiate themselves.

I might get more of that together over the summer. Right now I'm just working out proficiencies and figuring out what I want to be part of the numerical bundle of the classes (I have only the sketchiest ideas for specific class features). The numerical parts of each class prevent things from getting super linear and allow for dipping to get low-level proficiency benefits, but maxing a proficiency is encouraged. If it ever becomes more than a way for me to waste time I should be spending on massive coding assignments, I'll post it here.

[BWR]

Feats are necessary for d20 in all its variants. Remove them and you will be forced to make a replacement system to account for PrCs, and if you don't non-casters will be even less useful than ever before. Restricting all build options to classes would mean either giving the classes some sort of replacement that emulates what feats did or making each classe highly restrictive in what you could do with it. I suppose you could make it work without feats somehow but I can't imagine anyone would want to play the game then.
Yes, there are a lot of problems with the feat system, with bad design and poor wording and whatnot, but that's hardly a problem limited to feats. Advantage systems in other games aren't necessarily any better balanced.

Feats are not necessary in all systems, considering the number of systems that don't have them and get along just fine.

[Eldan]

I like feats as class-independent character features. They work quite well that way. But as has been mentioned, a lot of them could just go.

Some are pretty pointless anyway, like all those which give +2 to one specific check. Just kill them. Since I'm toying with the idea of making saves and base attack a form of skill check that is bought with skill points, that counts for weapon focus et al too.

Some are extensions of things you an already do. There's a few that can just be rolled into skills, as has been mentioned. Like Endurance.

Some should just be basic features. I don't think anything would break if Power Attack was a combat special maneuver like bull rush or grapple. Or if light weapons automatically had weapon finesse.

But not everything quite goes there. Whirlwind attack is theoretically quite cool (if not worth the prerequisites), not class focused and shouldn't just come automatically with higher base attack. Or a lot of magic feats. Those can stay as feats.

[JCarter426]

2. Paired advancement. This would sort of be like gestalting, except you'd gestalt between two mechanically very different aspects of your character. So for example you could level in 'training' on one side and 'persona' on the other side, or something like that.

Pros and Cons: Kind of a weird system, but combinatorically allows a reasonable degree of customization. Still very linear though.

I'm idlly working on a system that uses a variation on this alternative, and that's actually the main reason I started this thread. My idea is that each character would, essentially, be a gestalt of a bunch of things:

a) Race: perhaps 1/4 of a well designed D&D 3.5 class or thereabouts.
b) Class: perhaps 1/2-3/4 of a well designed D&D 3.5 class, also determines your access to
c) Proficiencies: Roughly equivalent to a skill + several feats/skill tricks, but covering much more ground as a whole. Each character advances something like 6-8 of these at once, and they're the main place where single-class characters get to differentiate themselves..
Just wanted to pop in to say that I'm idly working on something like this as well. "Idly working" - I like the sound of that.

I'm not sure if I'd keep feats, either. But I have divided up the progression almost exactly as you described.

First there is race, which follows a progression system rather than being a single choice at character creation. Racial progression could allow for more balance, I think, because not everything would have to be given at character creation. It would also offer more customization options. Rather than having dozens of elf subraces, you could just have the one race with different options at each level. And it would allow for more complex hybrid choices. For example, you might have a human character with traces of non-human blood from distant ancestors, reflected with a few levels of elf and a level of halfling, with the rest all human levels - similar to dipping.

Class progression would include class skills, spells, etc. - but not as much would be tied to class as is currently. I haven't made any decisions yet, but likely hit points and base attack bonus would not be tied to the class. One option I've considered would be to make some classes cost more XP to level up.

The third type is anything that isn't inherently tied to either of the above, such as feats and attribute points; you keep getting those at a regular pace, regardless of your other choices. You might get extras due to class or race, but everybody still gets the same batch of stuff here. There are also a few elements from d20 Modern that I'd like to slip in, such as character skills and starting occupations.

So in short, I'd say the issue is more about what specialization is available only at character creation. In the current system, a lot of choices are made at character creation and are irrevocable. Feats are one of the mechanics that fill that gap. But if you change the system so those areas are more flexible, then there isn't as much of a gap, and you might be able to drop something else.

[unbeliever536]

I like feats as class-independent character features. They work quite well that way. But as has been mentioned, a lot of them could just go.

Some are pretty pointless anyway, like all those which give +2 to one specific check. Just kill them. Since I'm toying with the idea of making saves and base attack a form of skill check that is bought with skill points, that counts for weapon focus et al too.

Some are extensions of things you an already do. There's a few that can just be rolled into skills, as has been mentioned. Like Endurance.

Some should just be basic features. I don't think anything would break if Power Attack was a combat special maneuver like bull rush or grapple. Or if light weapons automatically had weapon finesse.

But not everything quite goes there. Whirlwind attack is theoretically quite cool (if not worth the prerequisites), not class focused and shouldn't just come automatically with higher base attack. Or a lot of magic feats. Those can stay as feats.

Why not make Whirlwind Attack something just a regular combat maneuver, or at least easily emulated with such? It's pretty close to a full attack, since you can already target each enemy within your reach with a different iterative. The only requirement would be that you have enough iteratives to hit everyone and that they be accurate. It's a moderately common houserule that iteratives get full or near full BAB already (or, possibly, there's one major brewer who uses that rule and I've seen him mention it a bunch), and it's not like anything would break if warrior types got a few extra attacks. (Seriously, does anyone think FoB is broken?). If you want to "Whirlwind Attack", you just make an attack against each foe within range as part of a full attack, and throw on a bunch of riders like knockback or a trip or something because you're a mid-high level mundane, and Mundanes Can Totally Have Nice Things (at least in theoretical homebrew. Theorybrew?). Or, if you're not in a situation where that's called for (which won't be that often, cool as it is), you just use your awesome attack routine on single foes in sequence.

I don't think you should have to spend one of a few very limitted resources for the sole purpose of being able to make a unique type of of attack that is rarely needed and barely an improvement over a regular attack routine.

As to magic feats, any in particular? I'm for opening item crafting to everyone (with appropriate investment in Craft skills, if GMs want to use whatever craft subsystem is provided), and metamagic I could see as, in various cases, unique to particular magic user classes, generally granted to everyone who is skilled in magic, or superfluous. Of course, I don't know what magic will look like if/when I'm done, and it will certainly be much different from what it is now. I'm not sure how much metamagic will even make sense in the final system.

What other "magic" feats are there that don't make sense as part of "skilled in <particular type of magic>"?

@JCarter:
I wasn't thinking of allowing open multiclassing on the race side, although I would like to split abilities into ancestral and cultural, and allowing people to take cultural abilities associated with any culture that they had resided in long enough, as determined by the GM (with some guidelines, of course). I dunno.

[Perseus]

Really though, are they? On a design level?

If a system has good, modular classes with plenty of options (lol) and allows open multiclassing, what do feats add?

Large numbers of feats seem to be (1) worthless, (2) optional class features, (3) extremely limitted forms of multiclassing, or (4) some combination of the above. I feel like there's nothing feats do that couldn't be done better by something else in a class based system. More and more, I look at the entire feat subsystem as a really bad effort to respond to "advantage" subsystems in point-based levelless rpgs. If I were writing a d20 based system, I would dump feats entirely, absorbing them into other parts of the system designed to fit the level-based paradigm.

Feats are not necessarily needed in a game if you have the options like you suggest.

However the idea of feats (features) that anyone can pickup is well liked so getting rid of that system may be a bad idea.

Instead of small microfeats you should look into macrofeats. Macrofeats are easier to balance and there can be less of them.

The problem with D&D feats is that combat, skill, magic, and whatever else are all lumped into one pool. This means you must balance or attempt to balance a mundane feat with a magic feat... Which will be hard since mundane stuff is low fantasy whereas magic stuff is high fantasy. The benefit a magic feat (quicken spell) will out weigh the benefit of a mundane feat (power attack) based on the systems each one uses. (Most of the time)

[JCarter426]

I wasn't thinking of allowing open multiclassing on the race side, although I would like to split abilities into ancestral and cultural, and allowing people to take cultural abilities associated with any culture that they had resided in long enough, as determined by the GM (with some guidelines, of course). I dunno.
Cultural, interesting. I hadn't thought of that. It amounts to the same thing, though - stuff that you level up but isn't tied to your class... and therefore is a dimension of specialization that doesn't exist in the current system. And therefore maybe you don't need feats for it... unless you just use the feat system for that but add prerequisites, though that would add to the problem of everything being in the same pool.

[qwertyu63]

Really though, are they? On a design level?

Yes, you fool.


If a system has good, modular classes with plenty of options (lol) and allows open multiclassing, what do feats add?

More choice, which is always a good thing.


Large numbers of feats seem to be (1) worthless, (2) optional class features, (3) extremely limitted forms of multiclassing, or (4) some combination of the above. I feel like there's nothing feats do that couldn't be done better by something else in a class based system. More and more, I look at the entire feat subsystem as a really bad effort to respond to "advantage" subsystems in point-based levelless rpgs. If I were writing a d20 based system, I would dump feats entirely, absorbing them into other parts of the system designed to fit the level-based paradigm.

In order:
1: Then make the useless ones better.
2: Not a problem, this is a good thing.
3: Not a problem, this is a good thing.
4: See the previous 3 points.

[Eldan]

What other "magic" feats are there that don't make sense as part of "skilled in <particular type of magic>"?

Mostly metamagic type feats, yes. Maybe not hte purely numerical ones (Heighten, Maximize, etc.), those are a bit boring anyway. But fun stuff, like Explosive Spell, Invisible Spell, etc.

Whirlwind attack is still a bit better than a full attack, too, just in a very limited situation. If there's 5+ enemies, you normally won't have enough attacks to hit them all. And in most cases, focusing on one would be better anyway.


1: Then make the useless ones better.

I'd love to see how you make a feat that gives, say, +2 to appraise and decipher script useful.

[unbeliever536]

More choice, which is always a good thing.


Eh. If I'm feeling choice-starved, I'll play an all-pointbuy system. In such a system, a feat-like class of abilities is pretty much required, since there are no classes. Feats in D&D are an unnecessary concession to such systems (I contend). D&D, though, is based on gaining levels and improving your class, plus some unique flavor from race/stats. We add skills and multiclassing to broaden player options after character creation, both of which fit pretty easily into the "you gain levels and get better" paradigm. Feats, though? They don't fit that paradigm so easy. A proper class based system ought to be able to produce sufficient customization through race, class, and skills that extra, common, singleton abilities are unnecessary.

To your specific points:



In order:
1: Then make the useless ones better.
2: Not a problem, this is a good thing.
3: Not a problem, this is a good thing.
4: See the previous 3 points.

Well, actually:

1) That means, mostly, taking fluff and adding some new mechanical ability for each such feat, which adds a ton of work, or Increasing the Numbers yet again, adding to the already enourmous problem of numerical bloat in 3.x.
2) Optional class features should be optional class features, not some extra thing all on their own. The rogue class, on its own, should allow the player to build the "best" rogue possible for a given role (given static ability scores and race).
3) Should be superfluous given that the system allows for actual multiclassing. If characters need to pseudo-multiclass to branch out into other subsystems (or need help from outside their classes), it's a sign that the subsystems don't play together as well as they should, as I explained upthread. Ie, the existence (and need for) such feats is a sign that the multiclassing system is inadequate.
4) Seen 'em.


e:
Eldan-
Well, off the top of my head, Explosive Spell could definitely go to characters good at Evocation, and Invisible could go to those good at Illusions. There are a few that I can think of that don't quite have a place (Born of Three Thunders comes to mind, that's not really generic Evocation, and it's definitely not anything else), but I think those could be replaced by slightly more generic stuff + good RP.

You've caught me on the Whirlwind Attack thing in the specific, but I still think there's a way to do it without specifically taxing a character for the ability to attack every foe they can get to. I'm just not quite sure what that would look like.

e2:
Perseus-
I missed your comment the first time around, but I'm not sure how I want to respond to it quite yet.

[Eldan]

Eh. If I'm feeling choice-starved, I'll play an all-pointbuy system. In such a system, a feat-like class of abilities is pretty much required, since there are no classes. Feats in D&D are an unnecessary concession to such systems (I contend). D&D, though, is based on gaining levels and improving your class, plus some unique flavor from race/stats. We add skills and multiclassing to broaden player options after character creation, both of which fit pretty easily into the "you gain levels and get better" paradigm. Feats, though? They don't fit that paradigm so easy. A proper class based system ought to be able to produce sufficient customization through race, class, and skills that extra, common, singleton abilities are unnecessary.

There are abilities, though, that more than one class want to have. Think of Silent spell. Is that beguiler or wizard? Is Power Attack fighter or barbarian? Explosive spell, would that go to the warmage, evoker or sorcerer? Who gets weapon finesse, combat expertise or spring attack, the swashbuckler or the rogue?
Sure, we can argue if those should come automatically, but I think there will always be abilities that fit more than one class. Feats are good, for those.

[unbeliever536]

"Evoker", in my system, wouldn't be a class. You'd just have Warmages and Sorcerors (and other types of mages) who are good at Evocation. They're the ones who would get Explosive spell, if they specialize to a sufficient degree. Your expertise with the broad types of abilities you have (I'm calling these "proficiency schools") determines what you can do with them. These abilities are independent of class, except insofar as classes make different proficiencies available. Class features would be more generic with respect to proficiency schools, but specialized with respect to the class's role. The better you are with something, the more cool things you can do with it. Casting spells stealthily might wind up as part of a proficiency entirely seperate from magic, or part of the proficiency associated with magic that doesn't grant spells.

Or it might just go to the beguiler-type. Part of the problem with the Wizard is that it's overly broad*. The fighter would have the same problem, if it was any good. I'd rather avoid near-generic classes, since that gives more design space for class features. I'm sort of drifting off my original topic, but my ideal specific-to-general balance point would be somewhere between the core wizard and the school-specialist casters like Dread Necro and Beguiler.

*e: given a large number of classes. Wizard/Fighter/Thief/Priest works fine when those four classes are the only four classes. When you start talking about Knights and Barbarians and Evokers and Transmuters? Need more specific class features.

[Seerow]

Personally, I really dislike the way 3e handles feats, but think some universal character customization option is needed. For the game I've been working on when I have time for it, I'm planning on something like this:

-Feats become the name for martial character special abilities. So yes, fighters get lots of feats. These feats are all specifically designed for non-casters, utilizing a resource system shared by non-casters, and all give the character the ability to do things actively.

-Most of the passive/boring feats that people don't care so much about, but are still interesting for the purpose of character customization become perks. Perks are divided by attribute, with an extra list for 'universal' perks.

-Each time a character gains an odd attribute point above 10, they gain 1 perk. They can either take a perk related to the attribute that went up, or a universal perk. So a character starting with a stat spread of 15/14/13/12/11/10 would have 9 starting perks. This front loads the character customization some, and has the added benefit of making odd attributes very desirable, so you don't feel like you're just wasting points having 2 15s instead of a 16 and a 14.

-Active feats for casters get modified into new spells or class abilities. So something like metamagic might become a swift action spell, cast this spell as a swift action to make the next spell you cast maximized or whatever.

[Vadskye]

To both of you (and everyone else, but specifically addressing this issue): What if the former functionality of feats (that is, specific options not strictly associated with a particular class) were rolled into a different system, such as skills. You can see some of this in what some people (brewers and late designers alike) do with skill tricks in CS, or with skills generally in Tarkisflux's Tome of Prowess (http://www.dnd-wiki.org/wiki/Tome_of_Prowess_%283.5e_Sourcebook%29). Why shouldn't, say, the excellent athlete, who has invested heavily in the Climb, Jump, and Swim skills, naturally gain the kind of fitness and coordination necessary to execute the manuevers represented by the various running feats (and similar abilities related to his athleticism)? Basically, can we absorb the best parts of the feat system into a more streamlined system for outside-class advancement that better fits the level-based paradigm of D&D?
I'm familiar with the Tome of Prowess. But most feats can't be mapped to skills in a way that makes sense. Skills measure something different.

Now, to be clear, you don't need "feats". You can call them perks instead. You can change the method of feat/perk acquisition, like Seerow suggests. But at the end of the day, they're still feats.

(Seerow, I'm intrigued - it reminds me of the Scion "knacks" - but it sounds horribly front-loaded. I think that would only make sense in a system where ability scores increased much more often after level 1.)

[Seerow]

(Seerow, I'm intrigued - it reminds me of the Scion "knacks" - but it sounds horribly front-loaded. I think that would only make sense in a system where ability scores increased much more often after level 1.)


I agree. My system has ability scores increase a lot past level 1 (in exchange for no more magic items or wishes or whatever that boost them). It runs something like this (http://www.giantitp.com/forums/showthread.php?t=236454) (tweaked a bit more to fit the specifics of what I'm doing, but that's the general idea behind it). It's still pretty front loaded, but not as much as it would be just transplanting the perks into 3e.

[OzymandiasX]

Feats add a lot of customizability that would not otherwise be there.

Simply having a handful of class options doesn't help in this regard. Sure you may have 5 different ways a rogue can specialize, but feats turn that into dozens of additional (useful) ways to specialize.

I think feats are one of the best additions ever added to tabletop RPGs.

[unbeliever536]

Now, to be clear, you don't need "feats". You can call them perks instead. You can change the method of feat/perk acquisition, like Seerow suggests. But at the end of the day, they're still feats.


What do they model, though? It seems to me that the core thing that feats model is discrete horizontal growth. Designers can then twist that model around with prerequisites so that they become a (rough) model for discrete vertical growth, but I think that we are better served by a model that doesn't need to be twisted to model vertical growth, if that is what we seek to model (and I think it is, since we base everything off of a character's "level", a vertical characteristic).


I agree. My system has ability scores increase a lot past level 1 (in exchange for no more magic items or wishes or whatever that boost them). It runs something like this (http://www.giantitp.com/forums/showthread.php?t=236454) (tweaked a bit more to fit the specifics of what I'm doing, but that's the general idea behind it). It's still pretty front loaded, but not as much as it would be just transplanting the perks into 3e.

That's a pretty nice system you linked to. I thought I would have more to say about it, but that's it.

As to the perks, what do you have in the general list? What you do with the attribute list sounds analogous to what I'm proposing to do with generalized skills/proficiencies, though I'm trying to avoid strict "+X to Y" abilities in order to simplify and compress RNG.

[NichG]

Just because we have a strong vertical growth system doesn't mean that its not good to also have a form of horizontal growth in the system as well. That doesn't really follow (also, nothing wrong inherently with having multiple vertical or horizontal growth systems). Its just an issue of making sure you know what you're looking at and what each piece does for the whole. Feats are a little schizophrenic there, but then again so is almost everything else in 3.5ed D&D.

I mean, you could argue that having a single, centralized vertical growth mechanic allows you to accurately gauge the power level of characters with respect to challenges. But feats are actually only a tiny part of that uncertainty in 3.5ed - something like 'is his class Wizard or Druid or Rogue or Monk?' is generally the dominant effect. If you were going to try to do something to control all aspects of significant power variation in the system you'd certainly (eventually) want to look at feats, but if you aren't actually trying to achieve something like 'every LvX character is exactly as strong/viable as every other LvX character' then it seems like you risk sacrificing a lot of richness for tiny gains in balance/predictability.

[Seerow]

What do they model, though? It seems to me that the core thing that feats model is discrete horizontal growth. Designers can then twist that model around with prerequisites so that they become a (rough) model for discrete vertical growth, but I think that we are better served by a model that doesn't need to be twisted to model vertical growth, if that is what we seek to model (and I think it is, since we base everything off of a character's "level", a vertical characteristic).


Different characters are going to have different areas of focus. That is what you are modeling when you have feats, perks, whatever. This helps with the feel of character identity, so your Rogue isn't exactly the same as that other Rogue over there, you can actually be meaningfully different while still retaining the same base chasis.

A good example of this is the discussions that come up frequently over on the 3.5 Discussion board about what levels are "realistic". You can see there how even a commoner who only has 2-3 feats can invest those in different ways to model a ridiculous amount of diversity among normal people.



As to the perks, what do you have in the general list? What you do with the attribute list sounds analogous to what I'm proposing to do with generalized skills/proficiencies, though I'm trying to avoid strict "+X to Y" abilities in order to simplify and compress RNG.

The stat-related perks tend to be things like "Add this stat to something new (in place of the old stat if applicable)", "Gain a new use for this stat", "Improve a skill related to this stat", and so on. Universal perks are for things I couldn't really think of where else to work them in, and might yet go away if I find a better solution (right now it includes basically all of the minor feats that affect a class's mechanics. This is where stuff like Extra Rage/Extra Turning, Spell Focus, Weapon Focus, etc fell. Basically stuff that you might want to add to a class at a cost, but aren't really necessary to play and don't fit easily into the existing resource structures).

If you don't like having +X abilities that's fine. Personally, I think they have a place in the game, just not a place where they are competing against more active abilities, which is why I separated them out from those (with active abilities now being the realm of class abilities). Having a game where everyone is at exactly the same place on the RNG for everything is pretty boring. Seriously 5e's bounded accuracy is the reason I refuse to give two craps about it.

[Debihuman]

Really though, are they? On a design level?

I think that they are.


If a system has good, modular classes with plenty of options (lol) and allows open multiclassing, what do feats add?

No system is that modular except perhaps GURPS but it is percentile based rather than class based.


Large numbers of feats seem to be (1) worthless, (2) optional class features, (3) extremely limitted forms of multiclassing, or (4) some combination of the above. I feel like there's nothing feats do that couldn't be done better by something else in a class based system. More and more, I look at the entire feat subsystem as a really bad effort to respond to "advantage" subsystems in point-based levelless rpgs. If I were writing a d20 based system, I would dump feats entirely, absorbing them into other parts of the system designed to fit the level-based paradigm.

Then, I would say that 1. You don't use your feats properly. 2. Feats should not be class features unless you mean bonus feats. I generally dislike homebrews that turn specific class abilities into feats as that cheapens the classes. 3. What does multiclassing have to do with feats per se? 4.The class system is supposed to be a limiting factor. Otherwise you pretty much guarantee that all members of a given class will look exactly the same, unless you have have so many choices of special abilities that no class would be recognizable as such. In that case you might as well play GURPS.

Debby

[Vadskye]

I've never seen this board be so much in agreement. I'm impressed.