Any interest in a 2/3 spellcasting progression?

[clash]

So we have the EK and AT as 1/3 casters. We have the Paladin and Ranger as 1/2 casters but something that recently occurred to me is that you could just as easily have a class with a 2/3 caster progression. Do people find this interesting? Be a little more magically inclined than a ranger or Paladin but not quite a fullcaster. It would allow you to be more of a caster than a paladin but more of a martial than say a valor bard or bladesinger. It might also make room for something like invocations on a long rest casting class. Just some random thoughts. Let me know what you think.

[DivisibleByZero]

That's the role that an half caster Ranger or Paladin plays in 5e.

In 3e, Bards were a 2/3rd caster. They got level 6 spells, because 6 is 2/3rd of 9.
Half of 9 is 4.5, which is rounded up to 5, which is what we see in this edition. The difference of a single spell level is fairly insignificant.

Even the 1/3rd caster is not actually a 1/3rd caster, they're a 4/9th caster. If it were 1/3rd, they would stop at 3rd level spells. But they get 4th level spells.

So what we actually have is:
A 4/9th caster (44%), which we call a 1/3rd caster (33%)
A 5/9th caster (56%), which we call an half caster (50%)
And a full caster.

Any potential 2/3rd caster would probably actually end up being a 7/9th caster.

edit:
In thinking about it, they did it this way because of the Multiclassing rules.
AT and EK get 1 caster level for each 3 levels.
Rangers and Paladins get 1 caster level for each 2 levels.
How would you create Multiclassing rules for a 2/3rd caster that weren't convoluted and confusing? With this in mind, doing away with the 2/3rd and making it an half rounded up actually makes perfect sense.

[ZorroGames]

No, I personally would not be interested. But go ahead and develp it and see if it scratches any itches.

[clash]

That's the role that an half caster Ranger or Paladin plays in 5e.

In 3e, Bards were a 2/3rd caster. They got level 6 spells, because 6 is 2/3rd of 9.
Half of 9 is 4.5, which is rounded up to 5, which is what we see in this edition. The difference of a single spell level is fairly insignificant.

Even the 1/3rd caster is not actually a 1/3rd caster, they're a 4/9th caster. If it were 1/3rd, they would stop at 3rd level spells. But they get 4th level spells.

So what we actually have is:
A 4/9th caster (44%), which we call a 1/3rd caster (33%)
A 5/9th caster (56%), which we call an half caster (50%)
And a full caster.

Any potential 2/3rd caster would probably actually end up being a 7/9th caster.


Well slight correction to that as the fractions are calculated by caster class levels and it doesnt take all 20 levels to get 9th levels spells. A half caster has the exact same spell slots as a level 10 full caster. A third caster has the same as a level 7 full-caster so 7/20(35%) in this case. A 2/3 caster would have the spell slots of a level 13(65%) full-caster so you are correct that they would get 7th level spells.

[DivisibleByZero]

First of all, the level they attain the spells is irrelevant.
2/3 of 9 = 6 = highest spell level for a *true* 2/3 caster.

Secondly, go back and read my edit.

[LeonBH]

How would you create Multiclassing rules for a 2/3rd caster that weren't convoluted and confusing? With this in mind, doing away with the 2/3rd and making it an half rounded up actually makes perfect sense.

You would take the class and give it additional spell slots every 3 levels like AT and EK, but give them 2 caster levels worth of upgrades in spell progression.

When multiclassing, you multiply the caster level of that class by 0.67 and round down. Then you add it to the total caster levels.

[DivisibleByZero]

You would take the class and give it additional spell slots every 3 levels like AT and EK, but give them 2 caster levels worth of upgrades in spell progression.

When multiclassing, you multiply the caster level of that class by 0.67 and round down. Then you add it to the total caster levels.

It's funny because I asked how you'd do it without making it convoluted and confusing, and your answer was something which would be convoluted and confusing to many players.

[clash]

It is actually quite simple. Every 1.5 levels in the class count as 1 level on the spellcasting chart.

[LeonBH]

It's funny because I asked how you'd do it without making it convoluted and confusing, and your answer was something which would be convoluted and confusing to many players.

It's sad that many players can't multiply decimals.

[DivisibleByZero]

It is actually quite simple. Every 1.5 levels in the class count as 1 level on the spellcasting chart.

That isn't simple.
3-to-1 and 2-to-1 are simple.
1.5-to-1 would need a specific chart, and the ways that chart would interact with the other multiclas spellcasting comparisons would create ridiculous headaches and drama for a large portion of the player base.

They got rid of 2/3 for a very good reason. That reason was the new rules on multiclas spellcasters sharing slots.

Please edit to following to make it "simple" to add in a 2/3 caster.

Spell Slots. You determine your available spell slots by adding together all your levels in the bard, cleric, druid, sorcerer, and wizard classes, half your levels (rounded down) in the paladin and ranger classes, and a third of your fighter or rogue levels (rounded down) if you have the Eldritch Knight or the Arcane Trickster feature. Use this total to determine your spell slots by consulting the Multiclass Spellcaster table.

And I mean simple for the common player, not someone with a degree in math.
No.
3-to-1 is simple.
2-to-1 is simple.
1.5-to-1 is not simple for the every day average Joe.

I'm not saying that I couldn't do it. I'm saying it isn't simple for everyone, and simple for everyone is the goal. That's why they did away with 2/3 in favor of 1/2 and 1/3.

[LeonBH]

Please edit to following to make it "simple" to add in a 2/3 caster.

Spell Slots. You determine your available spell slots by adding together all your levels in the bard, cleric, druid, sorcerer, and wizard classes, two-thirds of your levels (rounded down) in the X class, half your levels (rounded down) in the paladin and ranger classes, and a third of your fighter or rogue levels (rounded down) if you have the Eldritch Knight or the Arcane Trickster feature. Use this total to determine your spell slots by consulting the Multiclass Spellcaster table.

I don't think you need a math degree to know how to multiply a fraction. And if you're multiclassing already, you're frankly taking more complexity than is necessary, and you can probably do simple multiplication.

[DivisibleByZero]

I don't think you need a math degree to know how to multiply a fraction. And if you're multiclassing already, you're frankly taking more complexity than is necessary, and you can probably do simple multiplication.
I repeat, this time with emphasis.

I'm not saying that I couldn't do it. I'm saying it isn't simple for everyone, and simple for everyone is the goal. That's why they did away with 2/3 in favor of 1/2 and 1/3.

[chokfull]

That isn't simple.
3-to-1 and 2-to-1 are simple.
1.5-to-1 would need a specific chart, and the ways that chart would interact with the other multiclas spellcasting comparisons would create ridiculous headaches and drama for a large portion of the player base.

They got rid of 2/3 for a very good reason. That reason was the new rules on multiclas spellcasters sharing slots.

Please edit to following to make it "simple" to add in a 2/3 caster.

Spell Slots. You determine your available spell slots by adding together all your levels in the bard, cleric, druid, sorcerer, and wizard classes, half your levels (rounded down) in the paladin and ranger classes, and a third of your fighter or rogue levels (rounded down) if you have the Eldritch Knight or the Arcane Trickster feature. Use this total to determine your spell slots by consulting the Multiclass Spellcaster table.

And I mean simple for the common player, not someone with a degree in math.
No.
3-to-1 is simple.
2-to-1 is simple.
1.5-to-1 is not simple for the every day average Joe.

I'm not saying that I couldn't do it. I'm saying it isn't simple for everyone, and simple for everyone is the goal. That's why they did away with 2/3 in favor of 1/2 and 1/3.

Boom. Now it's simple. You don't have to use decimals *in* the spell slots, you just have to match level 3 to level 2, level 6 to level 4, etc., all the way down the line. You can adjust them a bit if you want, but they look pretty much like this:

Lvl Slots
1: -
2: 2
3: 3
4: 4 2
5: 4 3
6: 4 3
7: 4 3 2
8: 4 3 3
9: 4 3 3
10: 4 3 3 1
11: 4 3 3 2
12: 4 3 3 2
13: 4 3 3 3
14: 4 3 3 3 1
15: 4 3 3 3 2
16: 4 3 3 3 2
17: 4 3 3 3 2 1
18: 4 3 3 3 2 1
19: 4 3 3 3 2 1
20: 4 3 3 3 2 1 1

And now that the table's been typed out, it's simple. No player has to get a degree in math to multiply decimals ever again.

[LeonBH]

No player has to get a degree in math to multiply decimals ever again.

I don't know about other schools, but you just need to go to first grade to learn that.


I repeat, this time with emphasis.

I also repeat:

Spell Slots. You determine your available spell slots by adding together all your levels in the bard, cleric, druid, sorcerer, and wizard classes, two-thirds of your levels (rounded down) in the X class, half your levels (rounded down) in the paladin and ranger classes, and a third of your fighter or rogue levels (rounded down) if you have the Eldritch Knight or the Arcane Trickster feature. Use this total to determine your spell slots by consulting the Multiclass Spellcaster table.

[DivisibleByZero]

Boom. Now it's simple. You don't have to use decimals *in* the spell slots, you just have to match level 3 to level 2, level 6 to level 4, etc., all the way down the line. You can adjust them a bit if you want, but they look pretty much like this:

You're all missing the point.
AT & EK: Every 3 levels in that class gets me 1 level on the chart.
Pally & Ranger: Every 2 levels in that class gets me 1 level on the chart.
Simple. No actual math required. Simple for everyone. That's the goal.

Your hypothetical 2/3 caster: Every 3 levels in the class gets me 2 levels on the chart. What about the levels in between? That requires math. Not everyone is good at math. Not simple for everyone, and the exact reason that 2/3 was removed in favor of 1/3 and 1/2.

Get it?

[LeonBH]

You're all missing the point.
AT & EK: Every 3 levels in that class gets me 1 level on the chart.
Pally & Ranger: Every 2 levels in that class gets me 1 level on the chart.
Simple. No actual math required. Simple for everyone. That's the goal.

Your hypothetical 2/3 caster: Every 3 levels in the class gets me 2 levels on the chart. What about the levels in between? That requires math. Not everyone is good at math. Not simple for everyone, and the exact reason that 2/3 was removed in favor of 1/3 and 1/2.

Get it?

Sure. I frankly don't care if the general populace can't multiply a whole number by two-thirds, and then round down. This is a homebrew question. Presumably, the OP can multiply.

[chokfull]

Your hypothetical 2/3 caster: Every 3 levels in the class gets me 2 levels on the chart. What about the levels in between? That requires math. Not everyone is good at math. Not simple for everyone, and the exact reason that 2/3 was removed in favor of 1/3 and 1/2.

Get it?

I just did it with literally no math. Levels 3, 6, 9, 12, 15, and 18 were simple. I just matched them to the level 2/3 below them, then progressed the spell levels naturally between. The natural spell progression is a pretty simple pattern. Besides, it's not hypothetical any more. I literally just made it. No one has to do math on it ever again, they can just copy the table.

[DivisibleByZero]

It's irrelevant.
We will never see a 2/3 caster in 5e, because they purposely removed it (from the previous editions where it existed) for a legitimate and logical reason. That reason being the new Multiclass rules.

You can make whatever homebrew you want, but we'll never see an official 2/3 caster. I guarantee it. They did away with it for specific logical reasons.

[LeonBH]

It's irrelevant.
We will never see a 2/3 caster in 5e, because they purposely removed it (from the previous editions where it existed) for a legitimate and logical reason. That reason being the new Multiclass rules.

You can make whatever homebrew you want, but we'll never see an official 2/3 caster. I guarantee it. They did away with it for specific logical reasons.

Cool. Yep.

[2D8HP]

So we have the EK and AT as...

.....Let me know what you think.


Sure, why not?

But the closest I've come to playing any kind of spell caster in 5e was a High Elf Rogue with the Firebolt Cantrip that I briefly played,

Otherwise every single one of my PC's have had bows and/or swords, and could sneak and perceive, I just find the magic rules to be too many options for me to keep track of and use well.

[Paeleus]

What class/character concepts do you hope to fulfill with a 2/3 caster?

Could the Warlock chassis be considered 5e's 2/3 caster?

[Eloel]

Your hypothetical 2/3 caster: Every 3 levels in the class gets me 2 levels on the chart. What about the levels in between? That requires math. Not everyone is good at math. Not simple for everyone, and the exact reason that 2/3 was removed in favor of 1/3 and 1/2.

Get it?

No. There's always going to be math in D&D. You add your proficiency bonus to your ability modifier. Sometimes you even add twice that. You calculate ability modifiers - by round-down division too.

If someone can't multiply an integer <= 20 with 2/3, they can't function in life, let alone D&D. It's fine.

[DivisibleByZero]

If someone can't multiply an integer <= 20 with 2/3, they can't function in life, let alone D&D. It's fine.

I'm glad that you feel superior to some people who might not have any particular aptitude in math. I don't presume to do the same.
I'm also glad that the designers took these people into consideration.

[clash]

What class/character concepts do you hope to fulfill with a 2/3 caster?

Could the Warlock chassis be considered 5e's 2/3 caster?

I could see the warlock being converted to a 2/3 caster but using normal spell slots instead of pact magic. I also tend to think of paladin and ranger spellcasting as a bit slow. I tend to do a lot of Ranger 2-5/Cleric x and paladin 2-5/fullspellcaster x multiclassing that could be satisfied by this type of progression instead.

[Tortferngatr]

Honestly, I'd be interested in an AT or EK variant that trades in some of their baseline class strengths (and archetype features from 50/50 multiclassing Wizard) for more consistent (2/3rds) spell progression and maintaining some other strengths/progression of the class. (Also, I would totally would love getting to use the Trickster capstone for spells higher than 4th level...)

Like...a Fighter or Rogue that wants to do more than merely *dabble* in the arcane arts like a standard EK or AT would, and is willing to sacrifice some of their martial potential to improve their casting (and wants to avoid awkwardness around cutoff points and multiclassing.)

For AT I'm specifically thinking something very vaguely akin to what Pathfinder seems to do with the Eldritch Scoundrel variant rogue, but perhaps as some kind of offshoot of starting Arcane Trickster? Maybe trading some sneak attack progression, later access to certain class features, and the level 10 ASI for extra spells from any school, and casting levels at 6, 9, 12, 15, 18 (or 5, 8, 11, 14, 17, 20 to hit 7th-level spells at 20 [and maybe Arcane Recovery/possibly a spellbook to fit the higher focus on spellcasting?])

(Don't ask me how multiclassing would work with them, though.)

[Vorpalchicken]

Why not take a full caster for 13 levels and a non caster for 7 levels? That's going to give you much nicer stuff than 20 levels of a 2/3 caster. Unless you have a hell of a capstone. (Not that capstones matter any way)

[clash]

Why not take a full caster for 13 levels and a non caster for 7 levels? That's going to give you much nicer stuff than 20 levels of a 2/3 caster. Unless you have a hell of a capstone. (Not that capstones matter any way)

Several reasons: Less even progression, either no extra attack at level 5 or no spellcasting at until after level 5. Want 1 class that realizes the concept. Any of the same reasons people take half-casters. Could be used as a subclass feature that upgrades spellcasting for rangers or paladins.

[Vorpalchicken]

A 2/3 caster should not get extra attack at level 5

[clash]

A 2/3 caster should not get extra attack at level 5
I dont really see why not. Some full spellcasters do.

[Vorpalchicken]

I dont really see why not. Some full spellcasters do.

If they devote their Archetype to being a second rate fighter, then I believe two classes can take it at 6. Hmm ok maybe extra attack at 5 wouldn't be so bad.

[clash]

Warlocks also get it at level 5 as a fullish caster. Granted it requires some investment, but as a 2/3 caster is already giving something up I dont see it as a problem.

[PeteNutButter]

I feel like a shaman type class might be a good fit for a 2/3 caster using Druid spell list, with maybe some “arcane” spells thrown on like domains tied to subclasses. They would have to be weaker on martial capability as to not over shadow the ranger and paladin.

As for them not doing it, because of the math, I think they could do it and just have a side box with a table showing the caster levels for MCing, just a chart with levels 1-19 and how many caster levels given per level. They have little side boxes explaining more complicated rules in lots of places in the published books.

[MrStabby]

I homebrewed a couple of 2/3 casters. A revised bard and a pale master class. It worked fine.

Like every other spell caster in the phb there was a table showing spells at each class level. Simple. I play with people able to look up numbers in a table.

It took a few cycles to make work - although the initial power was on the low side. You get one action per turn to do the most effective thing you can. Trading melee power for casting power usually means being worse at both - your best action in most circumstances is relatively poor.

Classes like ranger/paladin side step this by having the magic support combat - their most popular use of spell slots is buffs and, in the case of the paladin, smites. Bladelocks get round this by only having powerful spells (and even then doesn't really work that well for phb warlocks.

If you want to see what happens when you add magic to a class that doesn't use that magic to enhance their combat ability, look at the 4 elements monk.

My work around had two elements. Firstly build a class that had access to lots of niche spells. The perfect spell for the situation is awesome, even if a low level. Earthbind, spike growth, and other conditional spells are good for this.

The second approach was to enable attacking and spellcasting in one turn. This is similar to valor bards, EK and sorcerer quicken spell. It also touches on the paladin/ranger phenomenon of a lot of good bonus action spells.

[Idkwhatmyscreen]

Boom. Now it's simple. You don't have to use decimals *in* the spell slots, you just have to match level 3 to level 2, level 6 to level 4, etc., all the way down the line. You can adjust them a bit if you want, but they look pretty much like this:

Lvl Slots
1: -
2: 2
3: 3
4: 4 2
5: 4 3
6: 4 3
7: 4 3 2
8: 4 3 3
9: 4 3 3
10: 4 3 3 1
11: 4 3 3 2
12: 4 3 3 2
13: 4 3 3 3
14: 4 3 3 3 1
15: 4 3 3 3 2
16: 4 3 3 3 2
17: 4 3 3 3 2 1
18: 4 3 3 3 2 1
19: 4 3 3 3 2 1
20: 4 3 3 3 2 1 1

And now that the table's been typed out, it's simple. No player has to get a degree in math to multiply decimals ever again.

I think that this table is simpler. Each time you go up a level, you get one more slot for that group's level. The exception is level 15 as you remain at 2 Lv.5 slots and you only ever have one 6th and one 7th Lv. Slot
Lvl Slots
1: 1
2: 2
3: 3

4: 4 1
5: 4 2
6: 4 3

7: 4 3 1
8: 4 3 2
9: 4 3 3

10: 4 3 3 1
11: 4 3 3 2
12: 4 3 3 3

13: 4 3 3 3 1
14: 4 3 3 3 2
15: 4 3 3 3 2

16: 4 3 3 3 2 1
17: 4 3 3 3 2 1
18: 4 3 3 3 2 1

19: 4 3 3 3 2 1 1
20: 4 3 3 3 2 1 1

Bold levels indicate multi classing breakpoints. Multiple them by 2, then divide them by 3 and round up. Levels below the breaking points are one less for multi-classing purposes.

[LeonBH]

snip

I don't think a 2/3 caster should have the Spellcasting feature at level 1. It should start at level 2. Otherwise, it creates wonky effects on the multiclassing rules.

For example, a Paladin 13 has one 4th level spell slot. If they multiclass into a 2/3 caster under your rules, their caster level becomes [round down (13/2)]+[round down (1*2/3)] = 6. A 6th level caster has no 4th level spell slots.

However, if they got the spellcasting trait at level 2, this scenario where casters lose spell slots due to multiclassing becomes impossible.

[Nidgit]

1: -
2: 2
3: 3
4: 4 2
5: 4 3
6: 4 3
7: 4 3 2
8: 4 3 2
9: 4 3 3
10: 4 3 3 1
11: 4 3 3 2
12: 4 3 3 2
13: 4 3 3 3 1
14: 4 3 3 3 2
15: 4 3 3 3 2
16: 4 3 3 3 2 1
17: 4 3 3 3 2 1
18: 4 3 3 3 2 1
19: 4 3 3 3 2 1 1
20: 4 3 3 3 2 1 1

With class features at 1, 3, 5, 6, 9, 11, 14, 15, 17, 18, and 20. I'd probably designate subclass features as 3, 6, 15, and 18. Starting at 4th level, the class always has more slots than half casters and less than full. You'd probably offset those first levels by granting two cantrips.

The main problem would be a 4 level dip, since an ASI and 2nd level spells is a pretty darn good level.

[LeonBH]

The main problem would be a 4 level dip, since an ASI and 2nd level spells is a pretty darn good level.

Then why not move 2nd level spells to 5th level? Seems like a good way to encourage people to stay in the class, at least.

[PeteNutButter]

Then why not move 2nd level spells to 5th level? Seems like a good way to encourage people to stay in the class, at least.

You wouldn’t want to start moving things around, because then it’s not really a 2/3 caster. Half casters get 2nd level slots at 5. It’s not a problem that you get 2nd level spells and ASI at fourth since the 3rd level will be presumably weaker on the way there. All full casters give 2nd level spells and an ASI by 4th level.

[DivisibleByZero]

2/3 caster table (http://www.d20srd.org/srd/classes/bard.htm)

[LeonBH]

You wouldn’t want to start moving things around, because then it’s not really a 2/3 caster. Half casters get 2nd level slots at 5. It’s not a problem that you get 2nd level spells and ASI at fourth since the 3rd level will be presumably weaker on the way there. All full casters give 2nd level spells and an ASI by 4th level.

It's an odd number though. You can only precisely map the spell slot equivalent of 2/3 casters at level multiples of 3.

It requires a 3rd level caster to possess 2nd level spell slots. That means if you take their character level and multiply by 2/3, and if you then round down, you must end up with 3.

So here's the caster level equivalent of a 2/3 caster from levels 2 through 6:



Char Level
Caster Level


2
1


3
2


4
2


5
3


6
4




As you can see, they become 3rd level casters at level 5. It just so happens that it coincides with half casters getting their 2nd level spells at the same level.

[DivisibleByZero]

It's an odd number though. You can only precisely map the spell slot equivalent of 2/3 casters at level multiples of 3.

It requires a 3rd level caster to possess 2nd level spell slots. That means if you take their character level and multiply by 2/3, and if you then round down, you must end up with 3.

So here's the caster level equivalent of a 2/3 caster from levels 2 through 6:



Char Level

Caster Level


2
1


3
2


4
2


5
3


6
3




As you can see, they become 3rd level casters at level 5. It just so happens that it coincides with half casters getting their 2nd level spells at the same level.

....except that 2/3rds of 6 is not 3....
Not even if you round down. And if you *are* rounding down, then why do they eventually get 7th level spells, as described above?

[LeonBH]

Fixed, thanks.


And if you *are* rounding down, then why do they eventually get 7th level spells, as described above?

Because two-thirds of 20 is 13, and 13th level casters get access to 7th level spells.

[DivisibleByZero]

{Scrubbed}

[LeonBH]

{scrubbed}

Yep, it seems that I made a mistake. How nice of you to point it out.

[Eloel]

{scrubbed}

Right, because players will be homebrewing a class of 2/3 casting. They won't. They'll look at a table.

[DivisibleByZero]

{scrubbed}

[DivisibleByZero]

Right, because players will be homebrewing a class of 2/3 casting. They won't. They'll look at a table.

The problem isn't with a spellcasting table.
The problem is with the multiclassing rules.
5e removed 2/3 casters and replaced them with 1/3 and 1/2 for exactly this reason.

[Eloel]

The problem isn't with a spellcasting table.
The problem is with the multiclassing rules.
5e removed 2/3 casters and replaced them with 1/3 and 1/2 for exactly this reason.

Were you there? Did you talk to the designers? Did you confirm with them that this was exactly the reason? They think 1/3 is fine, but 2/3 is suddenly too hard? Are people unable to double a number?

Even in a game where you meet every week AND level up every session, people have a full week to calculate a number*2/3. Are you saying there are people that cannot do that? Do you really believe that?

[PeteNutButter]

Such hostility everywhere...

I think you’d want to give the spells on the earlier side. I’m AFB, but I believe that is how other classes do it. EKs get 3rd level slots at level 13, right? That’s not level 15 (5x3).

Paladins/rangers are the same way getting 3rd level spell slots at level 9, even though half of 9 rounded down is only 4, which wouldn’t give 3rd level slots.

Yeah, so they always give them a little sooner than the caster level for MCing would actually allow.

[dejarnjc]

It took a few cycles to make work - although the initial power was on the low side. You get one action per turn to do the most effective thing you can. Trading melee power for casting power usually means being worse at both - your best action in most circumstances is relatively poor.


I think this is an important point. For 2/3 spell casting class you really need a spell-list tailored to the class so you don't get stuck in situations where...

Round 1:
2/3 spell caster casts a 4th level spell, it's moderately effective
Full spell caster casts a 6th level spell, it's super effective
Full martial attacks, it's very effective

Round 2:
2/3 spell caster attacks, it's moderately effective
Full Spell Caster casts a 5th level spell, it's very effective
Full martial attacks, it's very effective

[clash]

For the purposes of continuing this discussion, lets just assume for a minute that people are either not multiclassing or can figure it out.

Putting that aside my thoughts for continuing development of this went 2 ways.

1) Have the main class fully fleshed out with some balance of martial and caster prowess, and subclasses add mostly theme/flavor
2) Have the main class only provide the structure of a 2/3 caster with the subclasses providing different compensations(martial, maybe totems, invocations, sneak attack progression?) to flesh out the class.

what do you think is a better approach?

[dejarnjc]

I think 1 would be easier to balance but 2 would be more interesting overall.

[Eloel]

For the purposes of continuing this discussion, lets just assume for a minute that people are either not multiclassing or can figure it out.

Thanks.


1) Have the main class fully fleshed out with some balance of martial and caster prowess, and subclasses add mostly theme/flavor
2) Have the main class only provide the structure of a 2/3 caster with the subclasses providing different compensations(martial, maybe totems, invocations, sneak attack progression?) to flesh out the class.

what do you think is a better approach?

I like something in-between a little better. Definitely not wildly different stuff like invocation vs totems, but maybe focusing on one of the 'main' class features, sort of like Land vs Moon druids.

[LeonBH]

Such hostility everywhere...

I think you’d want to give the spells on the earlier side. I’m AFB, but I believe that is how other classes do it. EKs get 3rd level slots at level 13, right? That’s not level 15 (5x3).

Paladins/rangers are the same way getting 3rd level spell slots at level 9, even though half of 9 rounded down is only 4, which wouldn’t give 3rd level slots.

Yeah, so they always give them a little sooner than the caster level for MCing would actually allow.

You're right. EK's get their 2nd level spells at 7th level, their 3rd level spells at 13th level, and their 4th level spells at 19th level. So without multiclassing, they get their spell slots one caster level earlier than if they multiclass.

Rangers, on the other hand, get their 2nd level spells at 5th level, their 3rd level spells at 9th level, their 4th level spells at 13th level, and their 5th level spells at 17th level. Seems to keep with the theme.

In that case, it does seem more appropriate to assign the 2nd level spells at level 4.


For the purposes of continuing this discussion, lets just assume for a minute that people are either not multiclassing or can figure it out.

Putting that aside my thoughts for continuing development of this went 2 ways.

1) Have the main class fully fleshed out with some balance of martial and caster prowess, and subclasses add mostly theme/flavor
2) Have the main class only provide the structure of a 2/3 caster with the subclasses providing different compensations(martial, maybe totems, invocations, sneak attack progression?) to flesh out the class.

what do you think is a better approach?

I favor 1, just because that's also how half casters and third-casters function. The base class is almost never just the spellcasting feature.

[clash]

Thats not a bad approach actually. I could see having a few weaker elements of the class that the subclasses make stronger.

[Eloel]

Thats not a bad approach actually. I could see having a few weaker elements of the class that the subclasses make stronger.

I don't know the flavor you're going for, but let's take Duskblade from 3.5 (it's an average gish)

You could have one that focuses on getting better at channeling stuff into people
You could have one that focuses on buffing yourself pre-combat
You could have one that focuses on being better at slashing people without magic
You could have one that lets you use spells that are maybe not directly combat related

[DivisibleByZero]

{Scrubbed}

[LeonBH]

{Scrubbed}

Look, I still don't care if people can't multiply a whole number by a fraction. You spotted the mistake and it was corrected, thus the mistake is gone. If you are going to hold that over my head, I'm happy to watch you repeat yourself.

[Aelyn]

{scrubbed}

[Talionis]

I think the idea of a 2/3 caster is no more difficult to multiclass than a Edritch Knight or an Arcane Trickster. Those are on 1/3 casting. So its doable.

Do you want to do it...

I think you'd need a reason to do it. It would need to fill a design space not already filled.

A class we don't have in the game, that might be able to use it is the Artificer. I did not understand why that class was a 1/3 caster. I also thought the pet should be a class in and of itself. A 2/3 caster might balance an Artificer... preventing it from access to the highest level spells but allowing it to be more a caster than a Ranger or Paladin, which seems reasonable. I'm not trying to balance it right here, only suggesting that I could see Artificier being a 2/3 caster being a good tool to balance Artificer.

[MrStabby]

I do think spell casting should come at 1st level. The hypothetical class is predominantly a caster. Casting spells is more its thing than fighting.

To throw something out there - if there is a different progression it could be made more different still.

Call it another type of casting - much like pact magic is, and in that same way it doesn't interact with normal spellcasting (multiclass complications resolved there). This means you could even do something like a half caster with more spell slots or still have the higher level spells and still have more spell slots. You could do a half caster and beef up the casting with something like arcane recovery (possibly even after EVERY short rest).

Is the 2/3rds caster because you would like to explore the design space between being a paladin/ranger and being a full caster (kind of like the warlock does) or is it to explicitly explore that particular progression?

[Grod_The_Giant]

Ehh... I dunno if there's really design room for it, is the problem. You've got full casters who acquire a fair bit of martial skill via subclass (Blades/Valor Bard, Bladesinger Wizard, Bladelocks, etc), and you've got full martials who acquire a fair bit of magical skill, either via subclass (EK/AT) or inherently (Paladin/Ranger). If there is space for a slightly-more-magical-than-a-half-caster, I'd say it's space best fitted to a Paladin or Ranger subclass that soups up their casting-- getting some cantrips, say, along with perhaps an Arcane Recovery type feature and some list expansion.

[Theodoxus]

1) Anyone complaining about the math obviously never dealt with BAB in 3.x and multiclassing - or THAC0...
B) A Warlock is essentially a 2/3 caster - they only get up to 5th level spells, and then 1 spell each of 6-9th that recharges on a long rest. I would suggest a simplified 2/3 caster follow that idea. If you really want to make them feel like 2/3 instead of full, grant them 2 spells of 6th and 7th and none higher. Boom done.
iii) Profit? maybe prophet... not sure.

[MrStabby]

Ehh... I dunno if there's really design room for it, is the problem. You've got full casters who acquire a fair bit of martial skill via subclass (Blades/Valor Bard, Bladesinger Wizard, Bladelocks, etc), and you've got full martials who acquire a fair bit of magical skill, either via subclass (EK/AT) or inherently (Paladin/Ranger). If there is space for a slightly-more-magical-than-a-half-caster, I'd say it's space best fitted to a Paladin or Ranger subclass that soups up their casting-- getting some cantrips, say, along with perhaps an Arcane Recovery type feature and some list expansion.

I don't know... level six spells play a bit differently to the others and I feel there is a qualitative difference made by getting access to them. Also I feel that an arcane "cleric" might be more the feel (with just a bit more combat power and a little less spells): primarily a caster but with access to heavy armour (from some domains) and extra melee damage from level 8 (again some domains).

I think a more spell focused paladin or ranger wouldn't quite fill the niche - both use spells to augment their combat ability. They don't use spells in place of combat ability. The cleric is ok-ish in melee to conserve spell slots. I guess I am saying that the 2/3rds caster concept is pretty dependant on the spell list.

[Asmotherion]

You can easily make a 2/3 caster by multiclassing.

A good example of it is the Sorcerer-Paladin, the Sorlock, or a multitude of Gish builds who take some non-Spellcaster levels and Sacrifice spellcasting progression in order to become better at Gishing. It's a perfectly good choice to make for a character if it's consious; Not all casters aspire to become Archmages after all.

[Theodoxus]

I don't know the flavor you're going for, but let's take Duskblade from 3.5 (it's an average gish)

You could have one that focuses on getting better at channeling stuff into people
You could have one that focuses on buffing yourself pre-combat
You could have one that focuses on being better at slashing people without magic
You could have one that lets you use spells that are maybe not directly combat related

So, Warlock, warlock, warlock or warlock... hmm... those of us voting warlock seem to be right...

[DivisibleByZero]

1) Anyone complaining about the math obviously never dealt with BAB in 3.x and multiclassing - or THAC0...

First, I played plenty of AD&D with THAC0.
Secondly, I'm not complaining about the math. I'm saying that it's more complicated than 5e wants or was designed for. 5e was designed for simplicity.
1-to-3 is simple. 1-to-2 is simple. Those are easy for everyone.
2-to-3 is not simple, especially when multicasting comes into it, and is far more prone to mistakes and errors. That would not be easy for everyone.
That point was driven home perfectly when one of the people claiming how easy it would be calculated it wrong the first time he broke it down. And that wasn't even a multiclassing breakdown, that was a simple table.

They removed 2/3rds casters and replaced with 1/2 and 1/3rds intentionally, because 2/3rds casters would make multiclassing spellcaster spell levels more complicated than 5e wants to be.

[Theodoxus]

Wow, you just won't let that die, will you? How do you know Leon didn't fat finger it?

Anyway, no comment on the warlock? Cool. cool. cool.

[Eloel]

Wow, you just won't let that die, will you? How do you know Leon didn't fat finger it?

That wouldn't fit the narrative, so it didn't happen. Easy.

[LeonBH]

I do think spell casting should come at 1st level. The hypothetical class is predominantly a caster. Casting spells is more its thing than fighting.

Well, once again, the problem comes in the multiclass rules. They become wonky if your caster level is not 1 or more if you take your class level and multiply it by two-thirds, because existing builds can lose spell slots as a result of multiclassing to this new class.

[Grod_The_Giant]

I don't know... level six spells play a bit differently to the others and I feel there is a qualitative difference made by getting access to them. Also I feel that an arcane "cleric" might be more the feel (with just a bit more combat power and a little less spells): primarily a caster but with access to heavy armour (from some domains) and extra melee damage from level 8 (again some domains).

I think a more spell focused paladin or ranger wouldn't quite fill the niche - both use spells to augment their combat ability. They don't use spells in place of combat ability. The cleric is ok-ish in melee to conserve spell slots. I guess I am saying that the 2/3rds caster concept is pretty dependant on the spell list.
I mean, the Paladin and Ranger aren't the exact right thing, but the general 5th-level-slots structure plus a bit of subclass seems sound. (Perhaps the oft-mentioned "arcane half caster?") Heck, you could even have the subclass come in at 1st level, Cleric-style, if you wanted to be sure of having a bit of magic at first level.

[Idkwhatmyscreen]

I don't think a 2/3 caster should have the Spellcasting feature at level 1. It should start at level 2. Otherwise, it creates wonky effects on the multiclassing rules.

For example, a Paladin 13 has one 4th level spell slot. If they multiclass into a 2/3 caster under your rules, their caster level becomes [round down (13/2)]+[round down (1*2/3)] = 6. A 6th level caster has no 4th level spell slots.

However, if they got the spellcasting trait at level 2, this scenario where casters lose spell slots due to multiclassing becomes impossible.

This is why I said that you should round up at level breaks.(and use the break -1 for levels below it) Making the progression 1,1,2,3,3,4,5,5,6,7... and so on. Each odd number occurs twice and each even number occurs once.

[LeonBH]

This is why I said that you should round up at level breaks.(and use the break -1 for levels below it) Making the progression 1,1,2,3,3,4,5,5,6,7... and so on. Each odd number occurs twice and each even number occurs once.

That's very non-standard though. It could work for a homebrew, of course, but it would be more in line with the rest of the casters if it just followed all the same rules.

It also would still cause people to lose spell slots if they multiclass to the 2/3 caster class, if they were already a Paladin 9 for example, and the 2/3 caster class was a caster at level 1.

[MrStabby]

Well, once again, the problem comes in the multiclass rules. They become wonky if your caster level is not 1 or more if you take your class level and multiply it by two-thirds, because existing builds can lose spell slots as a result of multiclassing to this new class.

Take 1, multiply it by 2/3 to get... 2/3, then round that up to 1. A single level of a 2/3 caster adds as much to total caster level as a level of a full caster.

[LeonBH]

Take 1, multiply it by 2/3 to get... 2/3, then round that up to 1. A single level of a 2/3 caster adds as much to total caster level as a level of a full caster.

Hmm, yeah, that doesn't make sense to me. It should be round down to keep it in line with other casters, who also round down. You either have to round up all the time, or round up only on certain levels. It's not very elegant.

[MrStabby]

Hmm, yeah, that doesn't make sense to me. It should be round down to keep it in line with other casters, who also round down. You either have to round up all the time, or round up only on certain levels. It's not very elegant.

Whats wrong with rounding to the nearest number? It seems to fit the profile and is simple as a rule. Other casters not doing it isn't a problem - other casters don't get 2/3rds spell progression either and we seem happy with that.

Certainly a rounding to nearest integer seems more reasonable/elegant than starting spell progression at level 2. Maybe this is a personal taste thing.

[LeonBH]

Whats wrong with rounding to the nearest number? It seems to fit the profile and is simple as a rule. Other casters not doing it isn't a problem - other casters don't get 2/3rds spell progression either and we seem happy with that.

Certainly a rounding to nearest integer seems more reasonable/elegant than starting spell progression at level 2. Maybe this is a personal taste thing.

Rounding is fine. Rounding up is an outlier though. You would be able to take two levels in this class and not lose any spellcasting progression as a full caster, which means its primed for a 2-level dip if the features are good enough at those levels.

It's inelegant because instead of writing homebrew that finds a way to follow the rules, it's just homebrew that makes a special exception of itself (more than what homebrew already does, of course). Multiclassing rules are not modular components, so changing that affects all the classes, instead of just brewing up one new class.

Starting the class at level 2 seems reasonable to me, given that half casters do that already, and this class hasn't earned their full 1st caster level by level 1 anyway. It's elegant if only because it works within the rules of the other casters (thus it can blend with them in terms of spell progression and not stick out as an obvious outlier), and it doesn't require changing the multiclass rules in the ugly (IMO) way that rounding yo the caster level does.

[Roland St. Jude]

Such hostility everywhere...Sheriff of Moddingham: Indeed. If people could dial it down in here, that'd be appreciated. Also, specifically, everybody drop the debate about whether the math is hard or not.

[Zalabim]

Probably too late for this thread, but there's already a progression that gets 2/3s of the spell slots of a full caster: Paladin and Ranger, with 15 spell slots at level 20 compared to 22 for a cleric. EK and AT get 11 spell slots by level 20, half that of a level 20 bard. I don't think "two more spell slots than the paladin"(from adding 6th and 7th levels) is enough design space to make a class. Instead, there's probably more that could be done with the warlock's style of getting 6th-9th level spells but not the full access to 6th-9th level spell scaling and selection. Basically, a half caster that gets powerful high level effects in place of traditional 6th+level spellcasting.


Rounding is fine. Rounding up is an outlier though. You would be able to take two levels in this class and not lose any spellcasting progression as a full caster, which means its primed for a 2-level dip if the features are good enough at those levels.

It's inelegant because instead of writing homebrew that finds a way to follow the rules, it's just homebrew that makes a special exception of itself (more than what homebrew already does, of course). Multiclassing rules are not modular components, so changing that affects all the classes, instead of just brewing up one new class.

Starting the class at level 2 seems reasonable to me, given that half casters do that already, and this class hasn't earned their full 1st caster level by level 1 anyway. It's elegant if only because it works within the rules of the other casters (thus it can blend with them in terms of spell progression and not stick out as an obvious outlier), and it doesn't require changing the multiclass rules in the ugly (IMO) way that rounding yo the caster level does.
It's just rounding normally instead of forcing to round up or down like most things do. Anyway, if a 2/3 caster gets spellcasting at the same level as a paladin or ranger, that's another point saying there's no space for 2/3 casting that isn't already filled by 1/2 casting.

[LeonBH]

It's just rounding normally instead of forcing to round up or down like most things do. Anyway, if a 2/3 caster gets spellcasting at the same level as a paladin or ranger, that's another point saying there's no space for 2/3 casting that isn't already filled by 1/2 casting.

Not necessarily, if the 2/3 caster gains higher level spells faster, and end up with higher level spell slots in the end. I think if you want a class that has spells at level 1, that should be assigned to a full caster.

[MrStabby]

Rounding is fine. Rounding up is an outlier though. You would be able to take two levels in this class and not lose any spellcasting progression as a full caster, which means its primed for a 2-level dip if the features are good enough at those levels.

It's inelegant because instead of writing homebrew that finds a way to follow the rules, it's just homebrew that makes a special exception of itself (more than what homebrew already does, of course). Multiclassing rules are not modular components, so changing that affects all the classes, instead of just brewing up one new class.

Starting the class at level 2 seems reasonable to me, given that half casters do that already, and this class hasn't earned their full 1st caster level by level 1 anyway. It's elegant if only because it works within the rules of the other casters (thus it can blend with them in terms of spell progression and not stick out as an obvious outlier), and it doesn't require changing the multiclass rules in the ugly (IMO) way that rounding yo the caster level does.

I don't think anyone takes a second level of a normal spellcaster for that extra spell slot: either 1 or 3 is more powerful where you go up in the level of spells you can cast. Even if you were to take a second level for that extra slot - so what? it isn't a problem. It is only a problem if the martial abilities are so strong that the combination of the two is too good.

Starting spellcasting from level 1 and using the intuitive rounding system seems to be the best option.

[BobZan]

I've made some two 2/3 casters and they worked wonders.

Both played the role of Archers.

1st: Fighter 1/War Cleric 2 (progress as you see fit)
2nd: Fighter 1/Divine Soul 2 (progress as you see fit)

[Lvl 2 Expert]

Alchemist in Pathfinder gets spells up to level 6.

I think the concept of a 2/3 caster would fit any version of d&d quite well and I'm kind of surprised we don't have more of them by now. Especially if you front-load the magic powers a little (rangers don't get cantrips and in 3.x even only started casting at level 4, a good 2/3 caster would pretty much be the opposite of that) you get a useful healer/buffer/controller/blaster/whatever the hall this class/archetype/build does but you don't run into the same late game problems a wizard or cleric gives you. Especially if you put a bit of a limiter on spell selection.

So yeah, I'm interested in these in general.

Of course you can always make one by just taking a non-caster level after every 3 caster levels, but it's not really the same is it?

[DivisibleByZero]

I don't think "two more spell slots than the paladin"(from adding 6th and 7th levels) is enough design space to make a class.

Precisely. Half casters replaced 2/3rds.

[LeonBH]

I don't think anyone takes a second level of a normal spellcaster for that extra spell slot: either 1 or 3 is more powerful where you go up in the level of spells you can cast. Even if you were to take a second level for that extra slot - so what? it isn't a problem. It is only a problem if the martial abilities are so strong that the combination of the two is too good.

Starting spellcasting from level 1 and using the intuitive rounding system seems to be the best option.

Well, a 2-level dip in Wizard is pretty nice. A 2-level dip in Cleric is justifiable. Jack of All Trades is decent for a 2-level dip in Bard. For non-full spellcasters, a 2-dip in Warlock is great. A 2-dip in Fighter is good.

Anyway, I didn't say people would take a 2-dip for the spell slots. Obviously they would do it for the class features (as otherwise, any other full caster class is competitive; it is the class feature that differentiates them). But in doing so, they would not lose any caster levels. That's not necessarily a problem, agreed. But it is worth pointing out.

Everything is rounded down in 5e. It feels very against the grain of the other casters to round up this caster's levels.

[Zalabim]

Everything is rounded down in 5e. It feels very against the grain of the other casters to round up this caster's levels.
The default is to round down if it isn't specified. However, a whole bunch of stuff about spellcasters and spell levels is either effectively or actually rounded up, like full, 1/2, and 1/3 casters access to new spell levels, ritual casting/book of ancient secrets learning higher level rituals, and Natural/Arcane Recovery's number of spell levels regained. Paladin and Ranger level 1, and then AT and EK level 1 and 2 are the only times spellcasters don't round up for spell access. Since a 2/3 progression is supposed to be faster than paladin/ranger's, it could start at level 1.

As another perspective on that, if you make a homebrew class that's supposed to have more emphasis on spells than paladins and rangers, but has for its spell chart
Level 1 (same as paladin)
Level 2 (same as paladin)
Level 3 (same as paladin)

you have failed in emphasizing spells. Those first one, two, and then three levels are highly important levels for establishing your class's identity, theme, and feel. The only thing I can suggest is giving the class cantrips or spell slots starting at level 1, and cantrips won't be satisfying since they can't overshadow simple weapon proficiency at 1st level. There just isn't room to separate the class's progression from a 1/2 caster any other way. You can't put it off and say the classes will diverge at level 7. The rounded down progression that starts at level 2 gives the 2/3 progression the same as the 1/2 progression for 4 out of the first 5 levels. Just skipping level 1 would have the class start getting spells at the level 2 part of a normal caster's progression (3 1st level spells). Starting the class with spell slots at level 1 makes the most sense.

So to give its progression in multiclassing, I'd give it the confusing method of "Add all of your levels in 1/1, 2/3 classes, half your levels in 1/2 classes, a third of your levels in 1/3 classes, then subtract a third of your levels in 2/3 classes (round down in all cases)," and fully expect people to miss either the first or the last clause including 2/3 classes (like how people don't initially notice that it doesn't include Warlock). I just don't know if it makes sense to make a class like this at all.

[LeonBH]

The default is to round down if it isn't specified. However, a whole bunch of stuff about spellcasters and spell levels is either effectively or actually rounded up, like full, 1/2, and 1/3 casters access to new spell levels, ritual casting/book of ancient secrets learning higher level rituals, and Natural/Arcane Recovery's number of spell levels regained. Paladin and Ranger level 1, and then AT and EK level 1 and 2 are the only times spellcasters don't round up for spell access. Since a 2/3 progression is supposed to be faster than paladin/ranger's, it could start at level 1.

As another perspective on that, if you make a homebrew class that's supposed to have more emphasis on spells than paladins and rangers, but has for its spell chart

Level 1 (same as paladin)
Level 2 (same as paladin)
Level 3 (same as paladin)

you have failed in emphasizing spells. Those first one, two, and then three levels are highly important levels for establishing your class's identity, theme, and feel. The only thing I can suggest is giving the class cantrips or spell slots starting at level 1, and cantrips won't be satisfying since they can't overshadow simple weapon proficiency at 1st level. There just isn't room to separate the class's progression from a 1/2 caster any other way. You can't put it off and say the classes will diverge at level 7. The rounded down progression that starts at level 2 gives the 2/3 progression the same as the 1/2 progression for 4 out of the first 5 levels. Just skipping level 1 would have the class start getting spells at the level 2 part of a normal caster's progression (3 1st level spells). Starting the class with spell slots at level 1 makes the most sense.

So to give its progression in multiclassing, I'd give it the confusing method of "Add all of your levels in 1/1, 2/3 classes, half your levels in 1/2 classes, a third of your levels in 1/3 classes, then subtract a third of your levels in 2/3 classes (round down in all cases)," and fully expect people to miss either the first or the last clause including 2/3 classes (like how people don't initially notice that it doesn't include Warlock). I just don't know if it makes sense to make a class like this at all.

You're right, there are lots of instances where the caster levels are given one level early - actually, all the half- and thirds- casters get their spell slots almost one caster level early, with the exception of their 1st level spell slots as you said.

Though, with that in mind, a 2/3 caster modeled after that pattern (where they get their next level spells one caster level early) will get their 2nd level spell slots at a caster level of 2+ instead of 3, and they will get their 3rd level spell slots at caster level 4+ instead of 5.

The reason why half casters have to wait until level 2 and thirds-casters have to wait until level 3 before getting the spellcasting feature is, I feel, because of the multiclassing rules, and because their caster levels are rounded down in the formula that combines all of the caster levels together in a uniform and non-confusing fashion.

So with all that in mind, the pattern seems to tell me that a 2/3 caster gets their 1st level spells at 2nd level, their 2nd level spells at 4th level, and their 3rd level spells at 7th level. The divergence with the half-casters happens at level 4.

Obviously, that doesn't fix the issue you raised where the first three levels are identical with the half-casters in terms of spell slot progression. I think there's pros and cons to granting the spellcasting trait at level 2.

The pro: if you grant the trait at level 2, you don't have to change the multiclass rules by much. It will still be, to add all your full caster levels, half your half-caster levels (round down), a third of your thirds-caster levels (round down), and now also two-thirds of your 2/3 caster levels (round down). It keeps the uniformity, which is great.

The con: what you said, wherein the 2/3 caster isn't differentiated from the half-casters early enough. They will have the same spell slots for the first 3 levels.

If you grant the spellcasting feature at level 1, the pros and cons are just the reverse of the above. You get to differentiate the class from half casters early, but you mangle the simplicity of the multiclassing formula.

I feel like it is still more elegant to go with granting the feature at level 2, though, because the fix will remain in that class alone rather than affecting all of the multiclassing rules. For example, granting a 2/3 caster their own version of Arcane Recovery, will effectively make them one caster level higher than the half-casters, without actually giving them one extra caster level. Another way might be to give them a scaling X/day ability to cast a 2nd level spell at 3rd level without using spell slots, just like Warlock invocations, which would effectively make their caster progression the same as a full caster's, but without actually making them full casters.

In the end, I think solving the issue by keeping in the grain of current casters while differentiating with half-casters through class features is the way to go in creating a class with this spell progression.

[FinalDuck]

Rounding is fine. Rounding up is an outlier though. You would be able to take two levels in this class and not lose any spellcasting progression as a full caster, which means its primed for a 2-level dip if the features are good enough at those levels.

It's inelegant because instead of writing homebrew that finds a way to follow the rules, it's just homebrew that makes a special exception of itself (more than what homebrew already does, of course). Multiclassing rules are not modular components, so changing that affects all the classes, instead of just brewing up one new class.

Starting the class at level 2 seems reasonable to me, given that half casters do that already, and this class hasn't earned their full 1st caster level by level 1 anyway. It's elegant if only because it works within the rules of the other casters (thus it can blend with them in terms of spell progression and not stick out as an obvious outlier), and it doesn't require changing the multiclass rules in the ugly (IMO) way that rounding yo the caster level does.

I think the the two approaches are based on how the class progresses.

With existing Paladins, 1st level is all martial, with a touch of magic, but no actual spells or slots. Then 2nd level is spellcasting, with a touch of martial (smites). This kind of progression should definitely round down.

However, seeing as 2/3 casters are more caster than martial, it could make sense to put the spellcasting first, and have the martial features wait until third level, like an inverted Eldritch Knight. That would mean that rounding up would make sense, because while a two-level dip would give you as much spellcasting as a full caster, it wouldn't give you any martial features.