Power Attack: 4e and 3e.

[Psychic Robot]

In 3e, Power Attack lets you move a slider around for your character's attacks, taking heavy penalties for an increase in damage, but it doesn't scale with level.

In 4e, Power Attack has a set penalty with a set bonus damage value (shitty though it might be) that scales (crappily) with level. Assuming that this value were better scaled so that it could fight against the padded sumo...

Which has the preferable underlying mechanic?

[Ice9]

Well, 3E power attack sort of scaled by how many feats you spent on it - things like Leap Attack, Combat Brute, and Shock Trooper could make it several times more effective.

Aside from that, it depends on the rest of the system - 3E Power Attack works well in 3E, where because buffs/debuffs/circumstances you often end up with a big gap between attack bonus and AC, which Power Attack lets you capitalize on. In 4E, though, things stay a lot more on the RNG, which would make 3E-style Power Attack pretty useless.

[RandomCasualty2]

Well 3E power attack sort of scaled by level, because at high levels, the system was divergent so that a -5 to hit generally didn't mean as much (since your first attack was assumed to be an autohit most of the time).

Where in 4E, a -2 penalty to attack would always hurt you, sometimes in 3.5 you could afford to take a -10 to attacks, especially once you factored in ToB maneuvers, where you were just taking an attack at full BaB, which was almost surely a hit.

[JonSetanta]

3e scaling version with a cap.

[CatharzGodfoot]

It does totally depend on the system. If the game keeps things on the RNG, it shouldn't scale at all or slide very much.

The 3e Power Attack might do better with a number of presets with BAB prerequisites. At BAB 1, you could do -1/+3; At BAB 5, -2/6; BAB 10, -4/+12... Or something. I probably have the numbers wrong.

[SphereOfFeetMan]

Which has the preferable underlying mechanic?

Neither. Power Attack, while now seemingly an iconic ability, is a worthless mechanic. What does PA do? It allows you to lower your chance to hit, in order to inflict more damage. In other words, PA is a math engine whose purpose is to enable you to inflict more average damage. You have to keep track of your changing attack bonus, damage, and the opponents apparent current Ac. PA is also a universally selectable ability which disproportionately rewards different classes/builds with higher average damage values. In 3.5 2-handers used PA, and light weapon users didn't. In 3.x, PA encouraged people to go off the RNG into crazy land. Not only did 2-handers try to hit enemies on a 2, they tried to hit on a roll of their attack bonus minus their Bab on their worst attack. If the Dm wanted to challenge the 2-handers attack bonus with a high-Ac opponent, now the Rogue/Monk/etc could only hit on a 20 b/c they were off the RNG. In 4e, it has static penalty/bonuses which stupidly is more useful to low-damage builds than strikers b/c of the lessened hit probability.

Power attack is a shit mechanic that has no value. If you want an ability to give you more damage, just give it more damage, like a Rogue's sneak dice. 3e/4e style Power Attack should not exist in any game.

[Voss]

3rd is blatantly better. 3rd edition power attack can actually matter with the rest of the combat system, while the 4e power attack is *always bad*. You are always taking a -10% chance of hitting for bonus damage that is increasingly irrelevant. It does indeed scale with level. Unfortunately, the scale is inverted as you do extra damage thats a smaller percentage of the enemies hit points with each level you gain.

3rd edition power attack has its problems, not the least of which is that apparently D&D players have become increasingly stupid to the point that basic math on the fly is somehow hard, but at least you aren't spending a feat on something that actively hurts you.

[Psychic Robot]

Sphere: Why so rage?

[SphereOfFeetMan]

Sphere: Why so rage?

Mostly because PA has become an irrefutable fundamental game rule in people's minds, despite the fact that it is based upon poor game design.

[Tequila Sunrise]

Like others have said, it depends on the other game rules. If you've got foes like 3e oozes that have the same AC at CR 20 than CR 1 and more hit points than a celebrity with substance dependence, you want 3e style PA. Because that's the only way to defeat that foe within 10 rounds or less. On the other hand if your combatant stats are more consistent like 4e, you want 4e PA, because nobody's ever going to take a -10 to hit anyway.

TS

[JonSetanta]

Power attack is a shit mechanic that has no value. If you want an ability to give you more damage, just give it more damage, like a Rogue's sneak dice. 3e/4e style Power Attack should not exist in any game.

Hmmmmm I think I'll go ahead and second this.

Also note that the implementation should not be anything like Weapon Specialization but more like a boost to the damage-increasing aspect of the Strength score.
No stipulations.
Nothing but pure raw damage to melee attacks.
Final Destination.

[Koumei]

more like a boost to the damage-increasing aspect of the Strength score.

Do you mean like:

"Your Strength modifier counts as X points higher for the purpose of damage added to melee attacks. You still multiply it all by 0.5 if it's a light/secondary weapon, or 1.5 if it's a 2-handed weapon, the only natural attack you possess, or whatever."

or:

"When you would normally add half your strength bonus as damage to a melee attack, you may now add your entire strength bonus. If you would add your strength bonus, you now add one and a half times the bonus. If you would already add one and a half, you now add double the bonus."

?

Either one could work reasonably well on its own. Obviously, if WotC had thought of doing it this way, they'd add them both as feats, taking care not to balance them together, and "The standard PA build" would always take both.

[JonSetanta]

"Your Strength modifier counts as X points higher for the purpose of damage added to melee attacks. You still multiply it all by 0.5 if it's a light/secondary weapon, or 1.5 if it's a 2-handed weapon, the only natural attack you possess, or whatever."

Looks good. I was going to say "Add X (scaling value) damage with melee attacks for each limb used to hold the weapon, or the same damage with each natural weapon." but yours works better within mechanics in a traditional sense.

It could also be "Add X value to Strength as a Y bonus only for the purpose of dealing melee damage."

[Voss]

On the other hand if your combatant stats are more consistent like 4e, you want 4e PA, because nobody's ever going to take a -10 to hit anyway.

TS

No you don't. In 4e, you never want to give up a single + to hit, ever. The damage you do is increasingly trivial, but if you don't hit, you don't plaster on status effects, and that actually matters.
No one wants to take a -10 to hit, but no one with any brains wants to take a -2 or even -1 to hit, especially for the shitty (or even actively harmful) benefits offered by the crap equipment selection or shitty feats like this.

[Tequila Sunrise]

*Shrug* To each his own. I can understand the argument that PA and similar options should simply add damage without penalty, but then you end up with a half dozen feats that all do the same thing; add damage. If that's what you want, just have one option like 3e Weapon Spec/4e Weapon Focus and let it stack.

TS

[Voss]

Sigh. No, that isn't what I want- that isn't the point at all. I want feats that are actually meaningful. It isn't just that 4e power attack is bad, its that its actually a stupid concept, completely antithetical to how the system works and scales. Taking it actually hurts you- like the 3e Dodge feat, its a posterboy of absolutely failed design- its stupid because the designers utterly fail to understand the math behind the system.

Feats were a good concept in 3e, however they were implemented half-heartedly, rather badly for the most part, and without any understanding that they didn't balance at all. [Quicken spell = extra actions = win; dodge= set resources on fire for no benefit = suck]. For 4e, they simply decided that the feats = dodge = suck model was for some inexplicable reason the way to go. Coupled with their usual ineptitude at balance, it means that there are a handful of feats that are worth taking because they actually give you a benefit [apparently by mistake], a couple that are actually useful because they improve otherwise crappy class/race features [but which you can't actually care about if you aren't that class or race], and pages of page of wasted ink.

Its not that 'not enough damage' isn't a problem in 4e, it totally is. But half-assed feats don't fix the problem, which is system-wide, and a simple function of hit points being CON + X*level (possibly *2 or *4), while damage is Stat mod+magic bonus+ (1 to 8)d<whatever> (based on power level and type). The slopes of the equations go to very different values.

[Tequila Sunrise]

Sigh. No, that isn't what I want-

*SIGH* Whatever. My point is simply that a game doesn't need a half dozen different options that all do the same thing. If you end up with a bunch of feats that all add damage, whether 3e PA style or otherwise, just collapse them all into one stackable option.

TS

[RandomCasualty2]

*SIGH* Whatever. My point is simply that a game doesn't need a half dozen different options that all do the same thing. If you end up with a bunch of feats that all add damage, whether 3e PA style or otherwise, just collapse them all into one stackable option.

Yeah a lot of the problem with 3E is that it had too many options that did the same thing, which led to rocket launcher tag when you stacked them all up.

[Voss]

Sigh. No, that isn't what I want-

*SIGH* Whatever. My point is simply that a game doesn't need a half dozen different options that all do the same thing. If you end up with a bunch of feats that all add damage, whether 3e PA style or otherwise, just collapse them all into one stackable option.

TS

Well good. As long as you're aware that you aren't actually contributing to the discussion at hand.

[JonSetanta]

Yeah a lot of the problem with 3E is that it had too many options that did the same thing, which led to rocket launcher tag when you stacked them all up.

Typed bonuses with small variety of types prevents this easily, but 3e has like 30-something different flavors of "I defy the RNG".

[RandomCasualty2]

Typed bonuses with small variety of types prevents this easily, but 3e has like 30-something different flavors of "I defy the RNG".

And don't forget all the untyped bonuses either. 3E actually could have been relatively sound if they just buckled down on bonus types. One thing Pathfinder should actually do is say that untyped bonuses don't stack.

[JonSetanta]

And don't forget all the untyped bonuses either. 3E actually could have been relatively sound if they just buckled down on bonus types. One thing Pathfinder should actually do is say that untyped bonuses don't stack.

Welllll IMO shrinking bonus type variety and labeling everything works better. I recently found this old thread from years ago wherein the same subject is discussed, and very relevant. You might remember, RC (RC2!).
http://tgdmb.com/viewtopic.php?t=37971&start=0

Leaving untyped bonuses of any kind screams "CharOp me!"

[IGTN]

The idea that I'd been toying with for RNG bonuses (as opposed to things like damage bonuses, which seldom interact with the RNG and are supposed to overwhelm it in most cases; a good damage roll is supposed to make the concentration check to keep a spell impossible, for instance), is separating everything into four types:

Level-based: This keeps you on the RNG. Nothing really needs to be said about it; it's its own bonus type, and, naturally, you cannot use cheese to stack multiple of these.

Primary: This comes from your main piece of equipment or character ability. For attacks, this is the magic on your weapon (probably), for AC, it's your armor, for skills, it's your training. It probably caps somewhere low, like at +5; Primary bonuses from different sources don't stack. The maximum for this doesn't need to be explicitly spelled out in the rules, since you can't stack them.

Secondary: These are your tiny incremental bonuses. Secondary bonuses of the same type don't stack, and they don't apply at all if they're typed like your primary bonus. Shields might be a secondary bonus, or they might be added to armor to produce one primary bonus. They also have a maximum, probably also +5. These bonuses might also have conditions that can remove them (A dodge bonus that comes from a fixed version of the Dodge feat would fit here), encouraging you to stack them higher than you're allowed to.

Some items might give a choice between a primary and secondary bonus; for instance, a Cloak of Elvenkind might give a +4 primary or +2 secondary bonus to Stealth checks. If you're trained (giving, for instance, a +5 bonus), you take the secondary bonus, but if not you take the primary bonus. The trained person is still better (+7 vs +4), and can still stack bonuses higher (+10 instead of +9) and cheaper (they only need +3 more instead of +5 more from items).

Tertiary: These are bonuses you or your buddies spent combat actions on. Flanking gives a tertiary bonus, as would an ability that gives you a bonus to hit people who just hit you. These have no limit as to how high they can stack, but are bound by type-stacking rules and don't stack with like-type secondary or primary bonuses.

The results of magical attribute bonuses are usually secondary bonuses (that is, the enhancement bonus from Gauntlets of Ogre Power gives you a +1 secondary bonus to all strength rolls); sometimes they're tertiary, if the bonus source is tertiary. I'm not entirely sure if base abilities fit here as a separate category like level-based or if they should add secondary (or even primary) bonuses.

With a bonus limit of +10 from primary and secondary bonuses, then, assuming no tertiary bonuses, you can go from 50% to just barely off-RNG with respect to a completely un-equipped enemy. Tertiary bonuses mess with this, but, since they're given to you by tactics and spending actions, they only allow you to apply status effect: tossed off the RNG wrt X to people in battle; if controlled well, they don't let you do more than that (controlled badly would be able to accomplish this for all rolls by buffing before combat).

[ubernoob]

IGTN, that's a good idea. Mind if I steal it?

Aside: Is there a way to make a 6 stat/ 3 save system work for all attacks and not break the RNG? I remember frank talking about it, but for my project I just need to make it work with 6 stats and 3 defenses (with varied attributes to attack with, just gotta remember how to set that up).

[Manxome]

Bonus types have always struck me as kind of dirty. If people are allowed to have or not have any particular bonus without sucking, then why do you care whether someone walks in with two X bonuses instead of an X bonus and a Y bonus? And if you need types because you expect everyone to have every type, then why not just give everyone all of the bonuses they're required to have to be competitive, rather than making them add up small numbers from a bunch of sources?

Of course, adding on stacking restrictions ends up being the most expedient solution in many actual systems, because for one reason or another things are divided haphazardly or they don't scale elegantly and chopping off the most egregious offenders is easier than fixing the theoretical underpinnings. Especially in a tabletop game, where the math has to be simple, even at the expense of balance. But they always make me feel kind of unclean.

Aside: Is there a way to make a 6 stat/ 3 save system work for all attacks and not break the RNG? I remember frank talking about it, but for my project I just need to make it work with 6 stats and 3 defenses (with varied attributes to attack with, just gotta remember how to set that up).

Depends exactly what constraints you have in mind. As far as I can tell, Frank's thing about powers of two and prime numbers is completely incorrect and never had any meaningful evidence backing it up, so I suggest you ignore it; the precise number of stats doesn't seem to be particularly important.

If you do a simple system with linear point-buy attributes and each attack being linearly boosted by one attribute, then there will be significant advantage to be had by specializing in one attack type. So you'll either have to live with everyone being a specialist or change that formula somehow.

You should probably start by asking: what kinds of builds do you want to be effective? What should their respective advantages and disadvantages be? Once you know those things, you can determine whether any particular system meets your design goals, and particular mechanics may suggest themselves as being useful. Make sure to think about defenses and how they're determined, as well. Tying offenses to defenses is one traditional method for counterbalancing the tendency for offensive specialization, but if you want that to work with 6 attributes you'll need to have multi-attribute attack/defense types (or more complicated shenanigans).