log in or register to remove this ad

 

Blog (A5E) Class Balance In A5E: How Much Damage Should A Damage Dealer Deal?

In Level Up: Advanced Fifth Edition, we’re creating new incarnations of the 5e character classes. Before we build our new classes from the ground up, we need… a teardown of the originals to see how they work!

Our design goal is to produce characters of approximately the same power level as the ones in the Players Handbook. We’ll need to do some math to figure out the targets we’re shooting for. Before we crunch the numbers, though, let's talk about what we mean by power level.

World Power Level

First, let me say that we're quite happy to expand characters’ abilities when it comes to the social and exploration pillars of the game. Some classes need more expansion than others. Currently, the wizard has dozens of exploration spells: scrying, teleportation, Jump, Find Traps, and many more. The bard has the social pillar covered, with Friends, Glibness, charms of all kinds, and the Expertise class feature which allows her to double her proficiency bonus. The rogue has Expertise but can't compete with the bard's spell tricks. A good roleplayer can do a lot with a fighter, but the class features don't do a lot of the heavy lifting.

We aim for each character class, including the non-spellcasters, to gain unique, powerful non-combat mechanical elements that let them do things that no other class can do. Let the spellcasters be jealous for once.



Combat Power Level

When I talk in this article about preserving the game's current power level, what I really mean is that a party of Level Up characters won't overperform or underperform a standard D&D party in combat. That means that you can play any D&D adventure, official or third party, and get the level of combat challenge that its designers intended.

Most of a class's combat statistics are pretty easy to figure out: How many hit points can we expect a fighter to have at level 10? What's a monk's typical Armor Class at level 3? Harder to calculate, but no less important, is this: how much damage can a character dish out at a given level? Without that piece of information, we can't really balance the classes' combat effectiveness.

There are so many variables in calculating damage that completely answering this question may be impossible. But we've got to start somewhere.

Let’s start with some assumptions.

1. I'll benchmark each character of level X against an enemy monster of CR X. Without some sort of class-granted accuracy bonus, each character hits 60% of the time. (Character attack bonus and monster Armor Class tend to increase at roughly the same rate.) If a class feature grants extra accuracy or advantage, that needs to be factored that into their average damage per round. (A mere +1 bonus to hit can result in an 8% damage boost!)

2. I assume that every area attack hits two monsters.

3. I average a character's damage over the first three rounds of combat.

4. For my benchmarks, I built Players Handbook-only characters, and I leaned towards the simplest subclass available. When presented with a class option, I chose the bigger-damage option. For instance, I built a Great Weapon fighter instead of a Protection build. I didn’t use feats, since I’d like this test to focus on class damage, not on feat effectiveness.

Before we start crunching numbers, we have an important decision to make. Which fight shall we simulate: an easy battle in which the party is conserving their resources, or an all-out assault where the wizards are using their highest spell slots, the fighter is using Action Surge, and the paladin is smiting everything that moves?

Why not both? Some classes can go nova, throwing down a lot of damage in a big fight, and some classes do steady damage throughout multiple fights. We need to be able to account for both of these strengths. So for each of the classes I surveyed, I charted their "no-resources" damage (using only infinitely-repeatable attacks they can perform at will) and their "nova" damage (using up every spell slot and class feature in order to maximize the amount of damage that they can deal).

To start, I charted the four "basic" D&D classes: the fighter, wizard, cleric, and rogue, plus two more I was interested in: the paladin, which I've heard is overpowered in combat, and the ranger, about which I've heard the reverse.

Here's my chart, on which I track average damage per round for the six classes for levels one through twenty. The solid lines represent maximum nova damage, and the dotted lines represent at-will damage. The rogue only has one damage line, because it really has no limited nova powers.

classchart1.png

The first thing that jumps out at me is that most of the classes fall into one of two categories: high-nova/low-at will, or medium-nova/medium-at-will. The evoker wizard and life cleric can really lay down a lot of damage in a boss fight, but when they're not burning spell slots they plink away with low-damage cantrips. Meanwhile, the champion fighter is right down the strike zone on every pitch. It's always producing around the same amount of damage.

Overall, I like the design of these classes. If it was me, I'd differentiate cleric a bit more by having it do more at-will damage and less nova damage than the wizard, but that's just a minor quibble.

The next thing I notice is that people are right about the paladin and ranger. The nova paladin puts out almost twice as much damage as the nova ranger (and my ranger is trying hard, using bonus action spells every turn and Conjure Volley when it becomes available). And the paladin doesn't give up much to the ranger in any other category to make up for all that extra damage. The paladin's at-will damage is only a hair under the ranger's. The paladin has better armor, the same hit points and better healing abilities.

I know I'm cherry-picking a bit here since I've chosen classes I know to be badly balanced against each other, and I'm compounding this by sticking to the Players Handbook ranger when I know there are higher-damage options in Xanathar's Guide. Nevertheless, it's good to get a sense of what the combat-effectiveness extremes look like.

The last class I want to talk about here is the thief rogue. Since it doesn't have any nova capabilities, you can judge it as either an at-will or a nova attacker. As an at-will user, it's among the better ones, keeping pace with the champion fighter. But judged as a nova class, it's by far the worst. It gets left in the dust by the nova champion fighter. In fact, it gains a big edge over only one nova class—the ranger—and only at levels so high that they are seldom played.

It's worth noting that so far I've only graphed one subclass for each of the classes I've examined, and subclass can make a big difference. To illustrate that, here's the battlemaster fighter graphed onto the same chart.

classchart2.png

The battlemaster is a much better nova subclass than the champion! It almost challenges the paladin for the melee damage-per-round crown. If we accept the fighter as the "right down the middle" class who always produces medium damage, this widens the strike zone a great deal.

So now that we've squinted at some charts, what conclusions can we draw for our character class redesigns?

Lesson 1: Character damage increases linearly with level.

It's a bumpy ride along the way, especially at the first level of each tier (5, 11, and 17), but on the whole, the classes I've graphed do somewhere around 5 + level damage when not using any resources, and somewhere around 5 + (3.5 x level) when they're going all out.

More work is needed here. These patterns need to be borne out with an examination of the rest of the classes and subclasses, more sets of different assumptions (what if character level doesn't equal opponent CR? What if area attacks hit 4 enemies?), and, of course, double checking the math.

Lesson 2: We should try to stay true to the AGGREGATE average damage numbers instead of maintaining each class's current Damage Per Round.

I don't think there's anything sacred about the paladin being the best nova melee class and the backstabbing rogue underperforming everybody. I'd be happy to adjust the damage outputs of the specific classes to better match peoples' story expectations.

D&D doesn't need to be perfectly balanced - it's not a pvp game - but there shouldn't be classes that are much stronger or weaker in combat. Most peoples’ D&D games feature a fair amount of combat, and everyone deserves to have fun during that chunk of their week.

Lesson 3: Damage isn’t dealt in a spreadsheet.

This isn’t something I learned from this graphing exercise, but it’s a reminder not to take it too seriously. The circumstances of every battle are different. And that’s vital to remember when we’re designing class combat features. Depending on the location and the opponents, each class should have a chance to shine.

Wizards should excel against big groups of weak foes clustered within fireball range. Clerics should wreck undead. Rogues should deal the most damage when attacking from ambush. As for the rest of the classes… that’s where you come in.

For the people who have stuck with me through this long post, I have some questions for you. I'd love it if you posted your answers in the comments.
  • For each character class (or for a few classes that you have opinions about), what are the combat circumstances in which you'd expect them to excel?

  • Am I overthinking this? Does combat damage matter to you?

Continue reading...
 
Last edited by a moderator:
Pail Hughes

Comments

Gammadoodler

Explorer
Fighters hit 20 in their attack stat at level 6 instead of 8.

Everyone has 20 in their attack/main spellcasting DC stat at 8.

Few other stats impact DPR to a significant degree (hybrids who use both saves and attacks are an exception). But my Paladin, for example, roughly 90% of its damage was weapon based; I don't think there are many classes who split their damage evenly between save-based and attack-based.

Fighter extra ASIs mainly boost defence without feats (AC, HP). Rogue extra ASIs mainly boost utility and defence (HP, cha/int/wis).
I think it's possible to model an impact to offense from feats without modeling specific feats' impacts. I don't know about using different rates for different classes, but a bump in projected damage for the classes with more ASIs seems like a reasonable result of them having more ASIs to spend (some of which could be spent on feats).

Edit: specifically these ASIs could be spent on offense-boosting feats, whatever they are, in a way that would be incremental to a capped attack stat.
 
Last edited:

log in or register to remove this ad


clearstream

Be just and fear not...
Rogue dex is capped (at 8) before their extra ASIs kick in.

So their extra ASI (at 10) don't boost their damage. It boosts con/wis/int/cha or something.
Your assumption is that the Rogue hits 20 Dex at level 8?

That means the model should assume +15% / +3 at level 1. And then at the end of tier 1 they gain +5% / +1 damage; and again at the end of tier 2 they gain +5% / +1 damage.

Seeing as almost all play occurs in the first three tiers, I do not see why this should be excluded from the model especially if ignoring feats.
 

NotAYakk

Legend
Your assumption is that the Rogue hits 20 Dex at level 8?

That means the model should assume +15% / +3 at level 1. And then at the end of tier 1 they gain +5% / +1 damage; and again at the end of tier 2 they gain +5% / +1 damage.

Seeing as almost all play occurs in the first three tiers, I do not see why this should be excluded from the model especially if ignoring feats.
No, the OP's model assumes monster AC scales with player ATK and gets 60% hit chance regardless of level. You hit on a 9+

L1 have +5 ATK and face 14 AC foes
L4 have +6 ATK and face 15 AC foes
L5 have +7 ATK and face 16 AC foes
L8 have +8 ATK and face 17 AC foes
L9 have +9 ATK and face 18 AC foes
L13 have +10 ATK and face 19 AC foes
L17 have +11 ATK and face 20 AC foes

Similarly, monsters save on a 13+. So we have:

L1 have 13 DC and face foes with +0 save
L4 have 14 DC and face foes with +1 save
L5 have 15 DC and face foes with +2 save
L8 have 16 DC and face foes with +3 save
L9 have 17 DC and face foes with +4 save
L13 have 18 DC and face foes with +5 save
L17 have 19 DC and face foes with +6 save

Is this a perfect model? No. It is far from perfect? Also no.

Now, fighters do hit 20 in their ATK stat at level 6; but by level 8 everyone else catches up. For the level 6-7 window, fighters can be assumed to have 20 strength (or dex), while everyone else has to be presumed to have 18.

That will generate a small blip in the fighter's graph at level 6 and 7. But I wouldn't really care if the OP missed it, because it doesn't matter to the trends.

Remember, the goal here is to get ballparks of what PCs are capable of, so when doing new design you don't end up in the wrong ballpark.

I ran the numbers on a Rogue/Fighter/Paladin to see how far off the OP was. And the OP was pretty spot on for Rogue, and a bit under for BM Fighter and Paladin. But I did have to optimize hard to pull off the Paladin and BM numbers, so that is reasonable.
 

clearstream

Be just and fear not...
No, the OP's model assumes monster AC scales with player ATK and gets 60% hit chance regardless of level. You hit on a 9+

L1 have +5 ATK and face 14 AC foes
L4 have +6 ATK and face 15 AC foes
L5 have +7 ATK and face 16 AC foes
L8 have +8 ATK and face 17 AC foes
L9 have +9 ATK and face 18 AC foes
L13 have +10 ATK and face 19 AC foes
L17 have +11 ATK and face 20 AC foes

Similarly, monsters save on a 13+. So we have:

L1 have 13 DC and face foes with +0 save
L4 have 14 DC and face foes with +1 save
L5 have 15 DC and face foes with +2 save
L8 have 16 DC and face foes with +3 save
L9 have 17 DC and face foes with +4 save
L13 have 18 DC and face foes with +5 save
L17 have 19 DC and face foes with +6 save

Is this a perfect model? No. It is far from perfect? Also no.

Now, fighters do hit 20 in their ATK stat at level 6; but by level 8 everyone else catches up. For the level 6-7 window, fighters can be assumed to have 20 strength (or dex), while everyone else has to be presumed to have 18.

That will generate a small blip in the fighter's graph at level 6 and 7. But I wouldn't really care if the OP missed it, because it doesn't matter to the trends.

Remember, the goal here is to get ballparks of what PCs are capable of, so when doing new design you don't end up in the wrong ballpark.

I ran the numbers on a Rogue/Fighter/Paladin to see how far off the OP was. And the OP was pretty spot on for Rogue, and a bit under for BM Fighter and Paladin. But I did have to optimize hard to pull off the Paladin and BM numbers, so that is reasonable.
This seems like a reasonable look at it, and yet doesn't match at all well what I see in play. I suppose that is because the model assumes that PCs don't use feats, don't have access to magic items, and dislike buffs.

I'm kind of partly persuaded, and yet I think if the designers ignore those things then their mechanics will break when they come into my game.
 

tetrasodium

Hero
Supporter
This seems like a reasonable look at it, and yet doesn't match at all well what I see in play. I suppose that is because the model assumes that PCs don't use feats, don't have access to magic items, and dislike buffs.

I'm kind of partly persuaded, and yet I think if the designers ignore those things then their mechanics will break when they come into my game.
I think that the designers need to accept that the odd game going for grittier feel with no feats & no magic items is going for a grittier feel & needs to scale back the CRs or find some other means to bridge the gap rather than do what stock 5e does & completely trivialize stuff in games where the norm of feats & magic items are in use.
 

Gammadoodler

Explorer
Rogue dex is capped (at 8) before their extra ASIs kick in.

So their extra ASI (at 10) don't boost their damage. It boosts con/wis/int/cha or something.
I think there are a couple different ways of looking at this.
1. A scenario where feats (that boost offense) are not allowed, and are not planned to be included in the revised rules. In which case, what you say here is true and resulting damage comparisons under these assumptions are meaningful. Or..

2. A scenario where feats (that boost offense) are allowed and are planned to be included in the revised rules. In this case, the model should account for some impact to incremental offensive output to the capped stat values. Specifically, there should be some bump relative to other classes for fighters and rogues that have access to additional ASIs above the norms.

The goal isn't (or shouldn't be) an absolute baseline damage value. It should be a value that would be comparable with what the designers are intending to accomplish in the finished product.
 

clearstream

Be just and fear not...
I think there are a couple different ways of looking at this.
1. A scenario where feats (that boost offense) are not allowed, and are not planned to be included in the revised rules. In which case, what you say here is true and resulting damage comparisons under these assumptions are meaningful. Or..

2. A scenario where feats (that boost offense) are allowed and are planned to be included in the revised rules. In this case, the model should account for some impact to incremental offensive output to the capped stat values. Specifically, there should be some bump relative to other classes for fighters and rogues that have access to additional ASIs above the norms.

The goal isn't (or shouldn't be) an absolute baseline damage value. It should be a value that would be comparable with what the designers are intending to accomplish in the finished product.
Also worth thinking about is that if one low-balls the curve, the worst case is that the designers use values that are a little too low. That's surely better than using values that are too high!
 

CapnZapp

Legend
Without having delved into the details of the analysis, let me just express hope Level Up is designing for the extremes rather than the average.

That DPS comparisons really use what numbers that can be achieved rather than mediocre averages.

Level Up will lose a lot of its appeal if it's as easy to break as 5E. By that I mean that there are a few known ways to get stratospheric DPS. And not even that is the problem. The problem is that it makes it really difficult to consciously select any other build. The power differential is simply too great. This reduces the available options for any gamer with the slightest need to not make suboptimal choices.

The problem with 5E is that isn't really nearly as much variety as it seems to a casual observer.

And no, I do not want a game as locked down as Pathfinder 2 or 4th Edition. It's much better if you design cool classes that really work differently mechanically speaking. And then make sure to buff the inferior builds (and possibly nerf the bestest ones). Balance is important. But avoiding the crushing sense of sameyness and "your choices don't matter" that plagues 4E and PF2 is more important.

You don't have to build the balance into the core framework. You just need to spend a design pass at the end of creation to make sure options don't vary as much as they do in 5E.

Just like I hope Level Up will boost the spells that rate red or purple in the guides. Spells should not work the same, but the end output should be comparable, if you want casters to have a wide selection to choose between. That fire spells are always better than, say, acid spells is really getting old by now, after five editions of just the same. That doesn't mean acid spells should be remade as a long range spell to deal 8d6 damage in a 20 ft sphere.

It means the spell might keep doing damage over several rounds, or debilitate the target(s) in other ways, so a player might conclude the acid spell is better in some cases, while fire spells remain better in other cases. It means that a Black Dragon Sorcerer should not be clearly inferior to a Red Dragon Sorcerer for no good reason.
 

CapnZapp

Legend
Also worth thinking about is that if one low-balls the curve, the worst case is that the designers use values that are a little too low. That's surely better than using values that are too high!
The proper way to balance any game is to design for the extremes.

It's better if the developers mistakenly believe a build intended to be a top tier damage dealer does 110 DPR at some level, but in reality players struggle to get over 90 DPR (say). The worst case scenario here is that other builds get explored and preferred by the player base. This increases choice.

The opposite scenario, where the designers mistakenly think their build tops out at 110 DPR, but canny players get it to do maybe 140 DPR, is worse, since every other damage-dealing build is now clearly inferior. This reduces choice.

In the case of 5th Edition, the situation is excruciating. Some builds deal easily twice the damage, without any discernible drawbacks. (And I'm not comparing to obviously ineffective builds. I'm comparing to builds the game clearly expect to be viable. But how much must you love a particular build to justify being half the hero of someone better at charbuild?!) The devs create monsters for non-optimized builds with minimal crunch, making nearly every min-maxed build using the tools provided by the PHB run circles around monsters.

It's much easier to nerf an encounter than to meaningfully improve upon it. That is, if three Blorgons is too much for a newb party, or one where nobody cares about DPR, the GM can easily use two Blorgons. Or use half-strength Blorgons. Or use lower-levelled monsters. Or simply play the Blorgons ineffectively.

But if three Blorgons is a joke to a hardcore group of veteran minmaxers that love crunch, the GM can't easily fix that without rewriting the encounter from scratch. Just doubling the hp (or the number of monsters!) is deeply unsatisfying in a game already accused of making big bags of hp. Just arbitrarily increasing attack or defense values isn't a very good fix either. Switching out monsters for higher-levelled foes often rewrite the encounter, since now maybe flight or teleportation can short-circuit intended challenges. And if the party is already level 14 or so, even CR 20 monsters might not stand a chance. This significantly reduces choice. This significantly reduces the worth of 5th Edition.

Any group that might consider moving over to PF2 simply to get more crunch that work better (especially at high levels) will hopefully get a better option in Level Up! :)
 
Last edited:

clearstream

Be just and fear not...
Without having delved into the details of the analysis, let me just express hope Level Up is designing for the extremes rather than the average.

That DPS comparisons really use what numbers that can be achieved rather than mediocre averages.

Level Up will lose a lot of its appeal if it's as easy to break as 5E. By that I mean that there are a few known ways to get stratospheric DPS. And not even that is the problem. The problem is that it makes it really difficult to consciously select any other build. The power differential is simply too great. This reduces the available options for any gamer with the slightest need to not make suboptimal choices.

The problem with 5E is that isn't really nearly as much variety as it seems to a casual observer.

And no, I do not want a game as locked down as Pathfinder 2 or 4th Edition. It's much better if you design cool classes that really work differently mechanically speaking. And then make sure to buff the inferior builds (and possibly nerf the bestest ones). Balance is important. But avoiding the crushing sense of sameyness and "your choices don't matter" that plagues 4E and PF2 is more important.

You don't have to build the balance into the core framework. You just need to spend a design pass at the end of creation to make sure options don't vary as much as they do in 5E.

Just like I hope Level Up will boost the spells that rate red or purple in the guides. Spells should not work the same, but the end output should be comparable, if you want casters to have a wide selection to choose between. That fire spells are always better than, say, acid spells is really getting old by now, after five editions of just the same. That doesn't mean acid spells should be remade as a long range spell to deal 8d6 damage in a 20 ft sphere.

It means the spell might keep doing damage over several rounds, or debilitate the target(s) in other ways, so a player might conclude the acid spell is better in some cases, while fire spells remain better in other cases. It means that a Black Dragon Sorcerer should not be clearly inferior to a Red Dragon Sorcerer for no good reason.
Well, this is also true. The risk with low-balling is to offer choices that are not choices, because they are overshadowed. I started out fearing that, then I thought about power-creep and considered if maybe there was safety in low-balling.

But no, I think LA has to offer powers that are competitive without being broken. For my tuppence the OP would retain their baseline model, but map another one onto it including foreseeable feats and magic items. GWM is foreseeable. A +1 to +2 attack/damage item in tiers 2 to 3 is too. Buffs are much harder to predict, but the main ones such as haste might be called out... then does the damage get attributed to the caster or the target?
 

CapnZapp

Legend
Well, this is also true. The risk with low-balling is to offer choices that are not choices, because they are overshadowed. I started out fearing that, then I thought about power-creep and considered if maybe there was safety in low-balling.

But no, I think LA has to offer powers that are competitive without being broken. For my tuppence the OP would retain their baseline model, but map another one onto it including foreseeable feats and magic items. GWM is foreseeable. A +1 to +2 attack/damage item in tiers 2 to 3 is too. Buffs are much harder to predict, but the main ones such as haste might be called out... then does the damage get attributed to the caster or the target?
I might clarify I'm talking more the total impact of an entire build - the end result.

Not the attractiveness of individual options. Cheers.
 

COMING SOON: 5 Plug-In Settlements for your 5E Game

Advertisement1

Latest threads

COMING SOON: 5 Plug-In Settlements for your 5E Game

Advertisement2

Advertisement4

Top