The final word on DPR, feats and class balance

Tanin Wulf · Jun 18, 2018

Umbran said:
Um, note that this is not factually correct.

Rock-Paper-Scissors is a perfectly balanced game. It does happen to be exceedingly dull, such that we don't play it as a game, but use it as a randomization method, but the point is still made.

The game is about guessing what your opponent will do. The player with the better intuition will have an advantage (yes, "perfect balance" must incorporate player skill into its effects in order to achieve perfect balance and parity between players; it's part of why it doesn't exist). In this situation, only two completely blind opponents, or two perfectly randomized solutions can produce a truly balanced result... which is not a game. (This then also suggests that any perfectly balanced game isn't actually a game anymore... yet another reason why it doesn't exist.)

EDIT: This is what I would also consider the defining line between balanced, and "perfectly balanced." These games you list are, indeed, BALANCED, in as much as any game probably ever could be.

Tic-Tac-Toe is a playable game, and perfectly balanced. It is also a solved game, such that no matter who starts, the game can *always* be forced into a draw.

Yes, it can, but notice in the solution: first turn ALWAYS has the advantage in Tic-tac-toe.

The card game of poker is perfectly balanced - the odds are the same for everyone.

Poker isn't about the cards, my good poster!

Tony Vargas · Jun 18, 2018

Tony Vargas said:
Balance is the maximizing of player choices, while keeping those choices both meaningful & viable.

For better or worse, that's the definition of balance I'll be using when I discuss it...

Tanin Wulf said:
Yes, it can, but notice in the solution: first turn ALWAYS has the advantage in Tic-tac-toe.

Since it's a solved game, we can know exactly how (im)balanced Tic-tac-toe is. It presents X with 9 initial choices, obviously. However, 6 of them are meaningless: it doesn't matter which corner or side you place your X in, the game's potential results set will be the same, regardless (there's nothing special about the top/bottom/left/right of the grid, it could be rotated with no effect on play). There's three meaningful choices for X: center, corner, or side. One of those, choosing a side square, is decidedly inferior to the others, the choice of corner square is clearly the best (again, it's a solved game, so we know that X starting in a corner & playing optimally leaves O only two paths to stalemate, both of which start with the center square). On his first move, O would seem to have 8 choices, but, if X has played corner & is playing optimally, has only one viable choice: center, the others result in certain victory for X.

If X is playing optimally, O can force a draw by playing optimally, but, if O makes one mistake, X can force a win. OTOH, if O is playing optimally, X must make two mistakes to allow O a chance at victory.

The game (rock-paper-scissors) is about guessing what your opponent will do. The player with the better intuition will have an advantage (yes, "perfect balance" must incorporate player skill into its effects in order to achieve perfect balance and parity between players; it's part of why it doesn't exist).

That is a very impractical definition of balance, since it relies on factors outside the game, itself, and yeah, that kind of balance would be problematic to hard-code into a system. However, many games do incorporate player ability to deliver /fairness/ (not balance, by the definition I prefer) - through various methods of handicapping.

Balance, however, as I see it, is about presenting players with choices, not about guaranteeing they make the right ones. So it's perfectly plausible for the better player to 'win' a balanced game quite consistently.

That said, draughts isn't exactly a very balanced game. It's devoid of trap or optimal choices, but, like utterly-imbalanced & solved tic-tac-toe, it presents a player with only 3 meaningful and viable choices. It's just that, it presents only 3 choices, period, not 9, 2 of which are meaningful & viable, then 8, only one of which is viable (and thus none are really meaningful). Arguably, there's no meaningful choice in draughts, either, so it fails that definition of balanced, entirely.

At that, I find myself tempted to do a very forge-like thing and create a completely unintuitive definition of 'imbalance' so I'm not going to. ;|

I'm going to leave it at neither game is an example of perfect, nor even of a non-trivial degree balance. One is perfectly fair, but very limited in total number of choices it delivers to players, the other is clearly imbalanced & solved, both are pretty boring.

Poker isn't about the cards, my good poster!

Again, like any competitive game it's more fair than it is balanced (I'm not deeply familiar, but the random aspect is presumably fair, but means players are presented with only a small sub-set of the game's possible choices with each hand - you can be dealt a 'bad hand,' too bad - I'd expect being dealer impacts 'balance' to some degree, etc). And, like draughts, it's much more about playing your opponents than playing just the mechanics of the game, itself.

Tanin Wulf said:
The reason why it's not a solved game at that point is because D&D has something the other games we're drawing comparisons to does not: the DM is changing the scenario. (Note this is not an appeal to Rule 0, it's an explanation of why it's not a solved game.)

I didn't mean to imply that it was, just that the charop meta-game, was analogous to 'solving' a sub-set of the game. Mainly in that there are sub-optimal & trap options that have been excluded from it. FWIW.

Your hyper optimized setup can be stopped, dead in its tracks, by being the wrong solution to any number of problems that are all perfectly valid expressions of the rules and the game and don't require houserules... merely the proper setup and execution from the DM. That's why it can't be a solved game and why a sub-optimal solution doesn't remove choices.

A sufficiently 'sub-optimal' build can very easily be strictly inferior, not only less effective at what it can do, but able to do less in terms of flexibility as well as power. If everyone has the same level of system mastery, that choice is de-facto removed from the chargen meta-game, if someone steps in the trap, his choices in play of the actual game are reduced.
Just imbalance in action.

(Noe that I'm not a fan of "traps" and would rather they not exist; I'm simply pointing out that this logic doesn't follow.)

Agreed on traps. I'm sorry I implied more than I intended in alluding to solved games...

Hussar · Jun 19, 2018

I'm frankly baffled how anyone can think that fighters are the DPS king of the fighty types. Every other class will outshine a fighter wihin its own niche. Rangers out damage all but the most twinked out fighters at range, barbarians out damage the two weapon fighters, and a sword and board fighter can't hold a candle to a paladin.

So, the sorcerer is out DPRing the fighter? Who cares? So is everyone else. Why shouldn't they get in on the action.

5ekyu · Jun 19, 2018

Tanin Wulf said:
The reason why it's not a solved game at that point is because D&D has something the other games we're drawing comparisons to does not: the DM is changing the scenario. (Note this is not an appeal to Rule 0, it's an explanation of why it's not a solved game.)

Your hyper optimized setup can be stopped, dead in its tracks, by being the wrong solution to any number of problems that are all perfectly valid expressions of the rules and the game and don't require houserules... merely the proper setup and execution from the DM. That's why it can't be a solved game and why a sub-optimal solution doesn't remove choices.

(Noe that I'm not a fan of "traps" and would rather they not exist; I'm simply pointing out that this logic doesn't follow.)

Enemy throws fog cloud - no advantage possible - the -5+10 output goes south...

Just one for instance.

5ekyu · Jun 19, 2018

Tony Vargas said:
Doesn't eliminate analysis as a useful tool. ...

That'd be testing playtesting, obviously, which is also helpful. Testing will find a problem, analysis will isolate the root cause, and point to possible solutions.

What we have, here, is a failure to communucate...

So, games present players with choices. The more choices at each choice point and the more choice points, the greater the potential depth of play.

But just adding choices doesn't always help. The classic example is the token in monopoly, you may like being the shoe, but it makes no difference in play. That's not 'meaningful' - RPGs add nuance to 'meaningful,' though (that you 'want to play a shoe' might carry some weight).
Another is the worthless choice - there's a variation of rock-paper-scissors that adds 'well' rock & scissors fall in the well, paper covers it - it obviates rock, so once both players realize that, the variation is back to three viable choices.

At absolute minimum, a viable choice must not have an alternative that is better than it in all ways. Again, RPGs add a lot of nuance to that.

Depends on how much heavy lifting that DPR has to do, in context (of the system), and what, if anything the other alternatives have going for them.

For instance in the assumed 6-8 encounter day with 5rnd encounters, 1500-2000 vs 1290-1670 vs 1050-1400 vs 600-1200. If for the sake of illustration, a profoundly simplified Mike Mearls style balancing of the games hypothetical full caster's slots, with cantrips filling in additional rounds, were equivalent to 1500-1650, then the 50 dpr martial balances at 6 encounters & is OP at 8, while the 43 dpr is below par at 6, but balanced at 8.
But, on a off-label 4 round day, the martials throw down 1000, 835, 700, & 400-600, while the caster, down 10 rounds of very hypothetical 15 dpr cantrips, is at 1350.

So it's not just "is it balanced?" In an RPG it's also balanced for what sort of campaign?

And, again, theres nuance. If you really like the style of a weapon that takes you down from 50 to 47.5 or 43 to 41.5, why not go for it? (Effing half-point on average differences.)

Of course, that's a D&Dish example, other games are less sensitive to day length.

Even on the OP board it doesn't mean that - optimization is a specialized exercise, it needs parameters. Usually optimized for a specific thing.

In an absolute sense, 'not strictly inferior,' should be viable, keeping in mind that relatively minor and highly situational qualities can save you from strict inferiority.

In an RPG, 'not consistently overshadowed in the scope of play' might be closer. Though, you'll note, thats a higher bar.

Balance is more important in an RPG, where play is ideally cooperative, and 'meaningful' can be independent of mechanics, than in the narrower scope of a board/video and/or competitive game.

Depends on day length... And pillar emphasis...

And that gets into another aspect - balance can be robust, or fragile...

Exactly. D&D traditionally copes with that by balancing to particular play expectations - a dungeon crawl with other adventurers waiting in the wings to jump your claim, new monsters moving in every day, old ones leaving with their hoards, &c; or 6-8 encounter/2-3 short rest days - it hasn't always been clear about those expectations or successful, of course.

One ed's lack of success at balancing classes led to sorting classes into Tiers by the power that mattered most in the highly variable context of an RPG: Versatility. Its still a useful tool to think about in 5e. Fighters lack versatility, but are solid tanks, Tier 4. Sorcerers have a potent spell list and cast spontaneously, but limited spells known that are hard to change, Tier 2. So, of course there's a corner case where the sorcerer can grind damage like the fighter, while in other scenarios going all in on some other spell.

Its important to remember that balanced doesn't mean identical. If all weapons do d6 (and no other qualities) theres no meaningful choice of weapon, if the wizard, sorcerer, and Psion all have the same slots, and identical spell lists, and trivial 'ribbon' class features, theres only one caster. Add or change something, give weapons different die types, proficiency, grits, damage types; give casters completely different spell lists, etc, and you avoid that, and re-eintroduce some balance, if you do it well.

A few points - too much shuffle to go at in depth..

Back on the statement that your definition is bringing flexibility vs restrictive into "balance" and your disagreement...

"What we have, here, is a failure to communucate..."

Your definition ssys balance is maximizing the options that are both meaningful and viable.

So from that ot would seem you see a game with 20 meaningful and viable options as more balanced (a higher maximum of options) than say a game with only 5 such choices.

That seems to be including flexibility in the definition of balance to me.

Is it not?

Second, i would tend to see the viable definition you present here - boils down to "better in some ways" as similar to what my stated goal as far as rpg "target as far as balance" goes... "Balanceable in play" (tho again to me the setting and types od challenges is the big honking 800lb gorilla in this.)

As for your encounter a day math, sorry but again the focus on dpr and total output fails to be convincing at all to me.

Whether someone wins or loses, succeeds or fails etc in six encounters (even combat encounters) is not determined by or even in my experience sttongly correlated to the estimated dpr of their attacks.

Thats because in those challenges the ability to do your white room sack of hit points output is a question with a plethora of trip wires.

Heavy armor guy or blurred grapples you, keeps hold, now you are unable to move, have just one guy to swing the axe at and - GWM 5/10 becomes useless.

Hold person, haste, slow, a lot of blindness, prone, etc options... Even going prone vs an enemy sharpshooter... All those turn the focus on white room dpr estimates on their head and the more and more one gets to the levels used to do this analysis in the micro - the more that focus in micro becomes more corrupting because it cuts more out.

As for balanced does not mean equal - exactly - at leadt until one starts using analysis of the micro to then insist on things like getting dpr outputs across martial to be the same.

That is the trap of micro-equality - insisting on getting the very small sub-parts equal - it does lead one towards equality in the form of identical.

The focus on "equilibrium" of the macro, the whole, allows stronger here but weaker there and puts "balanceability" at the front.

To me, the focus of rpg balance should center on showing and explaining when a is better than b, when c is better than a, etc.

As i have said, to me balance would be achieved if at the "big level option" each had uncommon circumstances where each option was "best", uncommons where they were "worst" and common cases where they were "average" - measured by contribution to success. That objective does not require micro-tuning outputs to within a few percentage points and can be directly tied in with "scenes in play" more than ever more excluding abstractions.

Tony Vargas · Jun 19, 2018

5ekyu said:
Your definition ssys balance is maximizing the options that are both meaningful and viable.

So from that ot would seem you see a game with 20 meaningful and viable options as more balanced (a higher maximum of options) than say a game with only 5 such choices.

That seems to be including flexibility in the definition of balance to me. Is it not?

Not the way I'm used to 'flexibility' being used around here, which tends to be about the range of capabilities of the character, no. I mean in terms of options presented to the player.

A game like Hero, for instance, is super-flexible, in part because it presents a finite set of options that are mixed & re-skinned as a matter of course to model virtually anything, it wouldn't get credit for 'infinite balance' just because each choice is so wildly flexible.

But if all you mean by flexibility is "presenting more options," sure.

Second, i would tend to see the viable definition you present here - boils down to "better in some ways" as similar to what my stated goal as far as rpg "target as far as balance" goes... "Balanceable in play" (tho again to me the setting and types od challenges is the big honking 800lb gorilla in this.)

I'd rather focused on balanced w/in the expected parameters of play. If the expected parameters are extremely permissive, balance is harder, if they're narrow (6-8 encounter days) it's easier. If a game is designed to narrow parameters, but in the field, gets used with broader ones, it's at best 'balanceable,' I suppose.

As for your encounter a day math, sorry but again the focus on dpr and total output fails to be convincing at all to me.

It was merely a hypothetical analysis along one dimension illustrating how viability could depend on factors outside the choices being compared, themselves.

WAs for balanced does not mean equal - exactly - at leadt until one starts using analysis of the micro to then insist on things like getting dpr outputs across martial to be the same.

Nod. It defeats the purpose: make two options identical, they're no longer meaningful alternatives to eachother. OTOH, 'better in some ways' often needs to mean 'closely comparable in critical ways' (in D&D, with it's race-to-0-hps combat dynamic, DPR is particularly critical), while meaningfully different in others. D&D happens to weight DPR pretty heavily, because merely 'wounding' an enemy does nothing. It's just an artifact of the system. A d20 game will be balanced more robustly if things like damage potential, bonuses relative to DCs, and resource pools are kept at a fairly close numeric parity - not because that's a desirable way to balance a system, but as a consequence of the most basic elements of the system - hps w/o any 'death spiral,' flat-distribution d20 resolution, etc...

The focus on "equilibrium" of the macro, the whole, allows stronger here but weaker there and puts "balanceability" at the front.

A problem with calibrating balance to the 'macro level' is that you end up with very limited options in how the game can be used. Fragile balance that only works when your campaign conforms to the macro parameters it's calibrated against.
When that balance is left out of the design phase, it's left to the GM to balance the game as he goes - and DMs have enough to do, already. So 'balanceable,' sure, is a thing (essentially "Imbalanced: please fix or repair daily"), so is 'fragile' balance ("OK! It's balanced! DON'T Touch ANYTHING!") - they're neither great things, resulting in games that take constant intervention to keep playable, or can be played only within a limited scope - but it's better than giving up, entirely.

dco · Jun 29, 2018

Tony Vargas said:
'Meh' on all counts. The hp advantage is 1/level. AC can be cheesed up in a variety of ways. IDK what BA has to do with healing, but Second Wind trails off in importance rapidly. And 'better' saving throws? The fighter is STR/CON, the sorcerer CON/CHA. In both cases, CON is the more important save, and CON a second-priority stat, while the less-important save maps to the primary stat (setting aside the otherwise very effective DEX-fighter builds, like the Archery SS). STR saves come up more often, but they're usually for comparatively minor effects, while CHA saves keep you from, for the major instance, being dominated (which, with a DPR build, is nice for your party).

Meh for you and that meh doesn't change the reality. The fighter without subclasses has exactly 2HP more on average, Second wind means more HP, they can have more AC, more ability score improvements can mean better saves using feats or improved abilities, Indomitable means rerolls for 1-3 failed saves depending on your lvl, ...

Yes, even the Sorcerer has considerable versatility. He /can/ put a lot towards DPR, or he can put it towards defense or whatever else he knows a spell for... he can't compete with the Wizard, but he's got it all over the fighter, that way.

What versatility? A few spells outside the combat focus?
Build an EK and that versatility difference goes down. If people use feats the fighter could use the 2 extra for 6 more skills/tools with proficiency, that's versatility for me.

No problem, compare that sorcerer to a SS fighter who can't be attack and is fully buffed and expending all his limited resources on DPR. Fair's fair.

As I said in an ideal situation is better, it is also a worthless situation in most cases, if you are not attacked you are probably not going to lose and DPR means nothing.

"Waste of time" I suspect. But, sure, maybe a Champion/Berserker or something? I mean, you are delaying or even losing extra attack by MCing out of fighter, and in a weapon-based DPR build that's not such a great idea, but MCing should certainly be open to both.

50% more rays if you use quickened spell once per short rest, if you want DPR...

--------------------------------

pemerton said:
Some posters seem to be arguing that a sorcerer is not comparable at DPR than a fighter because it is better off spending spell resources doing other more interesting stuff. I don't follow that argument. If a sorcerer can match, or come close to matching, a featless fighter in DPR, and is better off doing other more interesting stuff leaving the DPR to the fighter, that seems to show that a sorcerer is just better.

I understand the argument that a sorcerer (or archer) needs a melee fighter for defence, which gives the fighter a distinctive support role. But I don't understand this other argument.

It is a reality that classes are not only defined by DPR and that combat involves more things than DPR.
I've played a sorcerer and didn't need a fighter for defense.

pemerton · Jun 29, 2018

dco said:
It is a reality that classes are not only defined by DPR and that combat involves more things than DPR.
I've played a sorcerer and didn't need a fighter for defense.

And how does this tend to show the viability of fighters? They're not needed for DPR. They're not needed to defend "squishies". So what distinctive capability does a fighter bring to the table?

Tony Vargas · Jun 29, 2018

pemerton said:
And how does this tend to show the viability of fighters? They're not needed for DPR. They're not needed to defend "squishies". So what distinctive capability does a fighter bring to the table?

To step into the accostomed role of Devil's Advocate, yet again:

They provide a simple 'training wheels' class for newbies to use, until they gain enough player skills to graduate to a more advanced class, on their way to magic-user (or some other Tier 1 caster starting with 3.0), As Gygax Intended.
They enable rewards for system mastery by acting as a class version of a "Timmeh Card" with superficially good-looking numbers, and an attractive/familiar/relateable concept.
They uphold the D&D tradition of having a fighter class that is good for nothing much but fighting (as the name implies).
They are expendable. Like redshirts, only tougher, so you don't have to replace them too often, though they are high-maintenance (less so in 5e as they have HD & second wind).
They give the party someone to use fighter-only magic items.
They slice, they dice, they make mountains of jullien fries!

Oofta · Jun 29, 2018

pemerton said:
And how does this tend to show the viability of fighters? They're not needed for DPR. They're not needed to defend "squishies". So what distinctive capability does a fighter bring to the table?

Tortoise, meet hare. A fighter isn't the flashiest class, but round-in-round-out they do a decent amount of damage. It will probably vary by campaign but in the campaigns I've been involved in we were never guaranteed a 15 minute work day. So when the orcs are still pounding on the door and the wizard is down to cantrips, they're still ~~slogging~~ slaying away.

In addition, there's a fair amount of flexibility and utility depending on your build, especially if you allow feats. Besides, even though I'm a veteran player (aka old) there are times when I just want to shut off my brain, swing a sword and eat some popcorn. Fighters are great for that.

The final word on DPR, feats and class balance

Tanin Wulf

First Post

Tony Vargas

Legend

Hussar

Legend

5ekyu

Hero

5ekyu

Hero

Tony Vargas

Legend

dco

Guest

pemerton

Legend

Tony Vargas

Legend

Oofta

Legend

Similar Threads