D&D 5E Classes, and the structure of DPR

[clearstream]

So the next slice is offence - how a character contributes to DPR. Again, the yardstick is an ASI. Half an ASI is 1pt, a whole ASI is 2pts, and a double ASI is 4pts. Here is the table -

Notice the brighter red text

• On my first pass, it seemed to me that all classes follow the same steps - 5th and 11th - so that DPR in tier 2 is a meaningful step up from tier 1, and tier 3 is a meaningful step up from tier 2; and that is mostly true (10/12 classes)
• One exception is the rogue - whose constantly scaling sneak attack lands them in the same place, but without pronounced steps at the tier thresholds
• The other is the barbarian - whose intent seems to be that they deliver their DPR by fighting on relentlessly through damage - notice how little they invest in offence compared with sustain; their ability to sustain allows them to make choices in regard to weapons and feats that other martials might do less well with
• For two classes - monk and ranger - this table fails to really explain their step in DPR at 11th; the reason why will become clearer later

As I've several times noted, there are two fundamental approaches to DPR

• Some classes increase their DPR by increasing their number of attacks - barbarian, fighter, monk, paladin, ranger, warlock
• Others increase their DPR by increasing their damage per attack - bard, cleric, druid, rogue, sorcerer, wizard
• A few switch their approach, increasing number of attacks at 5th, and damage per attack at 11th - monk, paladin, maybe ranger
• The designers seem to consciously lean into diversity so that some classes get to make a choice - bards are an example - more on that later

There are two facets that I didn't capture, and perhaps should have. One is crowd control - effects that temporarily change how a creature participates in the fight. On the whole, control either adds sustain (controlled creatures don't attack) or offence (controlled creatures deal damage to one another). Another is buffing - effects that boost ally sustain or offence. Controls and buffs give characters a choice about how they want to approach a combat. Full casters - especially wizards - enjoy the greatest amount of such choice. Perhaps they should have their own table - control/buffs.

Finally, on investment into offence

• Barbarian, fighter and rogue invest remarkably little directly into offence - 14-16pts
• The full casters, having put less into sustain, invest greatly here - 42-52pts
• In the middle are monks, paladins, rangers and warlocks - 28-36pts
• Remember that all points are not equivalent: a few mechanically well-aligned features can have more impact in play than a plethora of poorly-aligned features; that's easily seen if you picture a character with 100 offensive features, of which they can use only one per turn...

Two posts to go, and then it might be possible to stitch it all together.

 

[clearstream]

Next explore, which captures features used outside of combat, on exploration and social mechanics. Again, the yardstick is an ASI. Half an ASI is 1pt, a whole ASI is 2pts, and a double ASI is 4pts. Here is the table -

What I found interesting here is how little even classes focused on these pillars invest in features for them

• The strongest three - bards, rangers and rogues - invest only 5-9pts (just a few ASIs)
• Clerics, druids, monks and perhaps paladins give a nod to these pillars - with 2-4pts
• The others make the minimum investment - 1pt - for their starting skills

Of course, casters generally have a few exploration and social spells they can cast, and sub-classes generally include a few features that support these pillars.

Next will the last table - sub-classes - and then I hope to look at a possible language for describing D&D classes. A set of terms that can be combined to express the design intent for each class.

 

[clearstream]

Finally, sub-classes, which for me fill in the blanks. Again, the yardstick is an ASI. Half an ASI is 1pt, a whole ASI is 2pts, and a double ASI is 4pts. Here is the table -

The analysis here isn't complete, but a few patterns already stand out

• The features of some classes - bard, cleric, druid, monk and ranger - can't be understood without their sub-classes, because they are entangled with them; for example, you cannot judge the full effects of bardic inspiration or ki until you read the sub-class features that draw on those pools
• Notice the bright red entries - monk and ranger get half their 11th level step from their sub-class; the design direction for monk sub-class features at that level seems to be to merge sustain with offence to try to deliver interesting ways to engage with combat, while for rangers it seems formed around an idea of attacking multiple foes, again creating (in theory) a distinct way to engage with combat
• Those 'd8-caster' classes - bard, cleric, druid and warlock - all gain a significant part of their offence through their sub-classes (druids also gain sustain, through necessary simplifications in the way wild shape works); the bard sub-classes that get extra attack at 6th level is an interesting example of how a choice can be offered
• Some classes - barbarians, fighters, paladins and rogues - pretty much do what they say on the tin; much of what they do is in the class itself, and the sub-classes are largely additive (paladins are a mild exception)
• Other classes - bard, cleric, druid, warlock, and I would argue monk and ranger - are much harder to assess on the surface; due to entanglements or commitments to sub-class features
• The fully committed casters - sorcerer and wizard - can't be assessed without knowing their spells, which is another dimension of the game altogether

Diversity abounds around a fairly small vocabulary of design patterns. One could call out battle masters as an example of fighters doing more than it says on the tin, and I would perhaps argue that battle masters are additive, and placed out in the fighter sub-class so that fighter can remain simple, while giving access to complexity for those who want it. Implying that the broader intent of fighters is the simplicity, not the complexity. I also have not done full service to features that are both sustain and offence, and especially not to control and buffing.

At bottom are the total class values at 11th level - converted back to ASI. I think the intent is that classes are worth 30-40 ASIs by that point. Remember that WotC designers might have used some other yardstick, and even if they did use ASIs, they might have evaluated features differently than I (something that input could be valuable in correcting.) I can't emphasise enough that characters seldom benefit from all their features, so fighters can be as effective with 28 ASIs as bards with 41, but bards will have more options.

This however probably reveals enough of the design patterns to make the next step, which is a design vocabulary for D&D classes.

 

[Cap'n Kobold]

So the next slice is offence - how a character contributes to DPR. Again, the yardstick is an ASI. Half an ASI is 1pt, a whole ASI is 2pts, and a double ASI is 4pts.

There are two facets that I didn't capture, and perhaps should have. One is crowd control - effects that temporarily change how a creature participates in the fight. On the whole, control either adds sustain (controlled creatures don't attack) or offence (controlled creatures deal damage to one another). Another is buffing - effects that boost ally sustain or offence. Controls and buffs give characters a choice about how they want to approach a combat. Full casters - especially wizards - enjoy the greatest amount of such choice. Perhaps they should have their own table - control/buffs.

You could just do a separate table for spells, since in most cases, they can count as Sustain, DPR, and exploration as well as buffing and crowd control. - Often with the class able to decide which they count as on an immediate or daily basis.

Finally, on investment into offence

• Barbarian, fighter and rogue invest remarkably little directly into offence - 14-16pts
• The full casters, having put less into sustain, invest greatly here - 42-52pts
• In the middle are monks, paladins, rangers and warlocks - 28-36pts
• Remember that all points are not equivalent: a few mechanically well-aligned features can have more impact in play than a plethora of poorly-aligned features; that's easily seen if you picture a character with 100 offensive features, of which they can use only one per turn...

Two posts to go, and then it might be possible to stitch it all together.

Could you do into a bit more detail about how spell access and slots are costed? Does a level one spell have the same value as a level 5 spell for example?

 

[clearstream]

Could you do into a bit more detail about how spell access and slots are costed? Does a level one spell have the same value as a level 5 spell for example?

In all cases, I considered class and subclass features against a) the value of increasing an ability score by 2, and b) reference feats. For spell access and slots, the reference feats are - drow high magic, fey touched, magic initiate, ritual caster, shadow touched, svirfneblin magic and wood elf magic.

When a class with spellcasting first gives access to a spell level, two slots come with that access, until spell level 4 when only 1 slot comes with access. By the level they gain access, they will have gained one or two more spells 'in-mind' (e.g. choosing Wis+Level to be on their cast-able list). I'm implicitly supposing that on average, one spell of fourth or higher is worth two of third or lower (because when you open a spell level, you get two slots if that level is third or lower, and only one if it is higher.)

Half of fey touched provides one set 2nd level spell and one 1st level spell chosen from a narrow list. Each spell can be cast once per day. Comparing with a 3rd level cleric, the cleric gains two 2nd level spells from a wide list. Each can be cast once a day. So the cleric spellcasting feature at that level is better than half of fey touched. Is it better than all of fey touched?

Drow high magic gives one set 2nd and one set 3rd level spell. Each can be cast once per day. It also gives a 1st level ritual spell at will (mildly better than just having that spell ordinarily.) Comparing with a 5th level cleric, the whole feat seems less good than getting two 3rd level slots and access to the whole list.

Regarding slots. A slot of a level you can currently cast I see as being worth not more than half of fey touched or drow high magic, which in effect give you two slots.

An argument could well be made that I am overvaluing access to a new level of spells. My thought was that cleric can make all of their spells in mind, 2nd level spells. To my observation, opening up a new level of spells is worth something in itself. So I think one has to value opening a new level of spells (and the spell slots and in mind that comes with) as worth more than a feat like drow high magic.

Those were the kinds of thoughts on my mind in making my assessment. What do you think? Any correction?

 

[Scribe]

I was out this weekend and missed this, I'll have to look it over.

 

[TheOneGargoyle]

In all cases, I considered class and subclass features against a) the value of increasing an ability score by 2, and b) reference feats. For spell access and slots, the reference feats are - drow high magic, fey touched, magic initiate, ritual caster, shadow touched, svirfneblin magic and wood elf magic.
<snip>
Those were the kinds of thoughts on my mind in making my assessment. What do you think? Any correction?

Intriguing work here. I'm keen to investigate this approach more.

Here are some thoughts on a few changes I would suggest:
Sustain:

• Armour - clerics get medium + shields which is a 1, fighters get heavy + shields which is 2, but barbarians get medium + shields, so this should be a 1, not a 2 ?
• Unarmored Defense - Monks UD is rated 2 but can't use shields, Barbarian UD is rated 1 but can also use shields with it. So these ratings should be reversed (Barbarian version is strictly better than the monk version) ?
• ASI - ASI's can be both sustain and offence, in my experience offence is taken far more often than sustain, so why do they go here and not offence ?

Net change: Barbarian +1-1=0, Monk -1

Offence:

• Rage - Barbarians get an extra rage at 3rd (should be +2 instead is not shown) & 6th (should be +2 instead of +1), and extra rage damage at 9th (should be +2 instead of +1) for a total of +4
• Fighter ASI's - at least 2 ASI's, probably realistically more like 4, will be spent on offence, so they should be at 4-8 pts higher esp if you incl GWM/SS (which should probably be rated as 4's each not 2's anyway), meaning Fighter could probably be up to +10 higher ?
• Ki - You have rated an extra sneak attack die (+1d6 dmg every round, so approx 35 dmg per short rest) as a 2 pt. 1 ki typically gives 1 attack once per short rest (so approx 8 dmg per short rest). So 1 ki doesn't seem worth 1 pt, it's probably more like 1/4 pt. So Monk should be -7 lower.
• Martial Arts - Increased MA die at 5th & 11th level add 1-2 dmg per round. Rogue Sneak Attack adds 3.5 dmg per round and is rated as 2 pts. So Increased MA die should be 1, not 2. So Monk should be -2 lower.

Net change: Barbarian+4, Fighter+10, Monk-9

Subclass:

• I'm not quite understanding how you have rated the subclass options - have you taken the strongest feature available from any of the subclasses ? Or an average across all subclasses ? eg, what makes the L6 subclass features for a Druid 1 pt, Barbarian 2pt and Bard 4 pt ?
• Where does the 11th level red subclass feature for monks & rangers come from ? Which subclass has been used for this ?

I must say I'm impressed at how much work & thought you've put into this

 

[clearstream]

Here are some thoughts on a few changes I would suggest:
Sustain:

• Armour - clerics get medium + shields which is a 1, fighters get heavy + shields which is 2, but barbarians get medium + shields, so this should be a 1, not a 2 ?
• Unarmored Defense - Monks UD is rated 2 but can't use shields, Barbarian UD is rated 1 but can also use shields with it. So these ratings should be reversed (Barbarian version is strictly better than the monk version) ?
• ASI - ASI's can be both sustain and offence, in my experience offence is taken far more often than sustain, so why do they go here and not offence ?

Net change: Barbarian +1-1=0, Monk -1

I agree on all points, and will adjust the tables.

Offence:

• Rage - Barbarians get an extra rage at 3rd (should be +2 instead is not shown) & 6th (should be +2 instead of +1), and extra rage damage at 9th (should be +2 instead of +1) for a total of +4
• Fighter ASI's - at least 2 ASI's, probably realistically more like 4, will be spent on offence, so they should be at 4-8 pts higher esp if you incl GWM/SS (which should probably be rated as 4's each not 2's anyway), meaning Fighter could probably be up to +10 higher ?
• Ki - You have rated an extra sneak attack die (+1d6 dmg every round, so approx 35 dmg per short rest) as a 2 pt. 1 ki typically gives 1 attack once per short rest (so approx 8 dmg per short rest). So 1 ki doesn't seem worth 1 pt, it's probably more like 1/4 pt. So Monk should be -7 lower.
• Martial Arts - Increased MA die at 5th & 11th level add 1-2 dmg per round. Rogue Sneak Attack adds 3.5 dmg per round and is rated as 2 pts. So Increased MA die should be 1, not 2. So Monk should be -2 lower.

Net change: Barbarian+4, Fighter+10, Monk-9

I think I agree on everything here, but just to check: are you saying you would rate each step improvement in rage as equal to an ASI? That doesn't seem right to me, as +2 on an ability score can give attack and damage, and other benefits. Example, the level 9 rage step gives half that: +1 damage.

Subclass:

• I'm not quite understanding how you have rated the subclass options - have you taken the strongest feature available from any of the subclasses ? Or an average across all subclasses ? eg, what makes the L6 subclass features for a Druid 1 pt, Barbarian 2pt and Bard 4 pt ?
• Where does the 11th level red subclass feature for monks & rangers come from ? Which subclass has been used for this ?

I must say I'm impressed at how much work & thought you've put into this

Thank you! My parsing of the class and subclass features for ranger and monks strongly suggests that their tier step up is gained via their subclass feature, as a mix of sustain and offence (for monks) or offence (for rangers). Take a look at the relevant levels and features and see what you think. An alternative theory would be that monks and rangers do not get a step up at the tier threshold, but I feel that isn't borne out by what you see in their class and subclass features, taken together.

EDIT Subclass features for levels aren't equal in power across classes, but they are across subclasses of a class.

 

[TheOneGargoyle]

I think I agree on everything here, but just to check: are you saying you would rate each step improvement in rage as equal to an ASI? That doesn't seem right to me, as +2 on an ability score can give attack and damage, and other benefits. Example, the level 9 rage step gives half that: +1 damage.

Actually that's a good point, I agree with you.

Bonus uses of Rage give +2 - +3 dmg at these levels, adv on Str checks & Str saves & resistance to wpn dmg. And given that we're only talking about 2-4 uses per day, getting an extra one is a big bump. So, yes, that's probably worth an ASI.

But extra Rage dmg is only going to increase their dmg by 1-2 per round, that's in-line with the increased Martial Arts die, so should be 1 as well.

So, net +3 for Barbarians not +4.

Thank you! My parsing of the class and subclass features for ranger and monks strongly suggests that their tier step up is gained via their subclass feature, as a mix of sustain and offence (for monks) or offence (for rangers). Take a look at the relevant levels and features and see what you think. An alternative theory would be that monks and rangers do not get a step up at the tier threshold, but I feel that isn't borne out by what you see in their class and subclass features, taken together.

Ok, I will take a more in-depth look and come back. Thanks.