2E multiclassing in 3E

[RobbyPants]

We all are aware of the problems of 3E multiclassing. I was thinking back to 2E, which seemed to give you a more genuine trade off.

In 2E you'd pick two classes and start out as level 1 in each. For things like Hit Dice, you'd roll HP and divide by 2 for each class. Back then, different classes advanced at different rates, and you had to track XP separately, but that wouldn't strictly be necessary for the concept to work. Because of the XP progression tables of 2E, it took double the XP for each subsequent level. Because your XP was being split between two classes, this effectively meant you were about a level behind everyone else. So, if the group was level 5, your Fighter/Cleric would probably be around 4/4.

That honestly didn't seem too bad. I'm not saying it's perfect, but you'd be advancing both classes at the same time, rather than one level at a time in 3E. The closest 3E got to this was classes like the Mystic Theruge, which was too little, too late. But, what if I took that basis and tried to make a PrC where you could stay just one level behind. This is a bit of a working prototype:

(Note: I was thinking this out with vanilla 3.5 rules, not Tome classes.)

---

Multi-class (Or, insert better name, here)

Prerequisites: At least one level in two different base classes.

Hit Die: An "average" of the two base classes. Add the number of faces on each die, divide by two, and round down to the nearest die type. So, a d10 and d4 = 10 + 4 = 14, divide by 2 = 7, round down to d6.

Base Attack Bonus: An average of the two base classes. I could write up a 7/8 and 5/8 progression table if needed, or we could round down to the existing three 1/1, 3/4, and 1/2 progressions. Personally, I'd prefer the first approach.

Saving Throws: Each saving throw is an average of the two base classes. Again, we could come up with an intermediate progression that is an average of the good and poor progressions. *

Skill Points: Get an average of the two classes. All skills from both classes are class skills.

Weapon/Armor Proficiencies: Nothing new gained.

Class Features: Each level, you gain all the class features of both base classes as though you'd gained a level in each.


* Also, I know every base class and PrC gives out that +2 bonus to every good save at level 1. If we wanted to pretend that this PrC wasn't actually a separate class, we could forego that bonus. I'm not heavily set on either approach.

---

So, how well would this work in practice? It's simple, and relatively easy to implement. I'm curious how abusable this is. The obvious benchmark is casters. They're what make the world go round. Doing this would cost you a caster level, which means you'd be behind in your highest level of spell available half the time, and half the time you wouldn't be.

The direct comparison would be the Mystic Theruge. At level 12, you'd be a Wizard 11/Cleric 11, which is pretty damn solid. That's a lot of spell slots, although, a Wizard 12 or Cleric 12 has quite a few by that point, too. Compared to the Mystic Theruge at level 12 would be casting at 9/9. That's clearly too weak, but is 11/11 too much? Probably your biggest gain would be at levels 3 through maybe 5 or 7. You'd get a lot of extra spell slots relative to what you'd normally have at that level straight-classed.

Similarly, you could take some martial class to beef up your chassis a bit at the cost of a caster level. You're probably better off just straight-classing.

Classes like Rogue could be thrown in for lots of Sneak Attack progression. I'm sure you could figure out ways to cheese it and get some pretty good damage progressions with touch spells. Still, I'm not sure it's that much worse than a straight caster.

Regarding martial classes, you could throw Rogue in and it'd be almost a straight-up improvement, at a slight cost to HP and BAB. I don't see this as a problem, as those classes get worse and worst as you gain levels, anyway.


Is there anything I'm missing? Obviously, some combinations fare better than others. Does this break the game terribly, or is it still the same Mystic Theruge problem, but just less so? I'd made something like this a while ago that required more sacrifice of class levels, but I wanted to try this 1/1 entry to see if mimicking a 2E multiclassing system was viable.

[Vaegrim]

For the most part, this just seems like a rediscovery of the Gestalt Characters variant. That variant straight-up confesses that Gestalt characters are stronger than the default and then tap-dances around how to handle it. Mucking with the XP charts to manage that balance seems like a straightforward answer, but I'm leery of it's effectiveness. Experience is a River seems to argue against this technique and I honestly didn't get enough experience with disparate XP rates during I played 3E/3.5 to get a feel for the impact.

[Omegonthesane]

Gestalt characters are straight up stronger when they are not one level behind the rest of the party on all of their class features. Having to be even one level behind is an incredible price for casters to pay however, which renders the comparison to gestalt flawed.

Mystic Theurge is actually a rather terrible choice of gestalt, because you don't get extra actions with which to use all those extra spells even before you get into the MAD (and no the Archivist list isn't good enough to justify SAD Wizard-Archivists)

[nockermensch]

A) Given the same levels of optimization, a level N gestalt character is patently weaker than 2 level N single-class characters. The gestalt has less HP, half the treasure and, most importantly, half the actions. Because the way challenge rating works, we know that two characters of level N are worth one character of level N+2. Therefore, the level adjustment for a gestalt is less than 2.

B) At the same time, also given the same levels of optimization, a level N gestalt character is patently stronger than 1 level N single-class character, because the gestalt is able to use the synergies of two sets of class skills. Therefore, the level adjustment for a gestalt is more than 0.

From A and B, we can conclude that the appropriate level adjustment for gestalt characters is 1.


TL;DR: Just let people play gestalt characters in normal games if that's what they want to do, but they pay 1 level for that. Don't let them use LA-buyoff to remove this LA, and the math just works.

[Omegonthesane]

A) Given the same levels of optimization, a level N gestalt character is patently weaker than 2 level N single-class characters. The gestalt has less HP, half the treasure and, most importantly, half the actions. Because the way challenge rating works, we know that two characters of level N are worth one character of level N+2. Therefore, the level adjustment for a gestalt is less than 2.

B) At the same time, also given the same levels of optimization, a level N gestalt character is patently stronger than 1 level N single-class character, because the gestalt is able to use the synergies of two sets of class skills. Therefore, the level adjustment for a gestalt is more than 0.

From A and B, we can conclude that the appropriate level adjustment for gestalt characters is 1.


TL;DR: Just let people play gestalt characters in normal games if that's what they want to do, but they pay 1 level for that. Don't let them use LA-buyoff to remove this LA, and the math just works.

Level adjustment is BS, simply having shitty hit dice for the number of real levels you should be losing out on means the gestalt doesn't die of Cloudkill.

(And getting back on topic the idea of a Multiclass class that gives gestalt features but rounded down averages instead of best-of saves and bab would then be less powerful than an equivalent gestalt of the same HD due to having to have two real hit dice each of which is only one class.)

[RobbyPants]

Gestalt is when you straight up have the best of everything. I’m talking about an average on the chassis. That’s how 2nd edition multiclassing worked.

[angelfromanotherpin]

Open multiclassing doesn't work. Gestalting doesn't work. This doesn't work, for the same reasons as gestalting. If you want combo-classes, just write combo-classes that meet the appropriate contribution bar.

[nockermensch]

Even when I just used LOGIC and FACTS to demonstrate that being a "straight best of everything" gestalt is also balanced by keeping the gestalt 1 level behind you insist on the weaker multiclass idea.

Which is fine, actually. If I had to write an actual complaint about 2e style multi-classing or gestalt is that these somehow increase the system mastery gap: You risk having situations where one dude rolls a wizard/warblade while other guy comes to the table with a fighter/monk. But assuming similar levels of optimization, these options probably increase the fun, specially if the group is small.

[jt]

Adjusting by one level is awkward for full casters. They get a power spike every other level when they get access to a new spell level, which means that how good a hybrid is would oscillate as they're caught up half the time. Some solutions would be to set them two levels behind (so they're always one whole spell level behind) or to somehow get them caught up to the whole level but hybridize the spells per day tables too (half each). On the other hand, back when the character optimization boards were active I proposed a prestige class where 3 levels each of two base classes would let you gestalt the rest of the way. The general consensus was that this was usually too weak, so 3 levels behind for hybrid instead of gestalt is definitely too weak.

The hybrid builds I'd watch out for are Any Full Caster / Factotum and Full Arcane Caster / Duskblade. Factotums can get extra actions and have mostly out-of-combat utility that's not constrained by the action economy. Duskblades can use a melee attack to deliver a touch spell, then later get the ability to multi-target a touch spell through a full attack, which can be used for all sorts of shenanigans if you're not limited to their garbage spell list. These hybrids are definitely more than one level better than standard classes, but if you're using 3.5 you're always going to get some builds that are way more powerful than others no matter what, that's the nature of the beast.

[RobbyPants]

Even when I just used LOGIC and FACTS to demonstrate that being a "straight best of everything" gestalt is also balanced by keeping the gestalt 1 level behind you insist on the weaker multiclass idea.

The problem isn't whether or not gestalt could work for this, or whether it's more powerful. I don't want a system where one player uses a variant system for some of their levels and not other levels, while another player doesn't use it as all. It's wonky as hell. This approach is just simply taking or not taking a PrC. That's something that already exists in the system.

Which is fine, actually. If I had to write an actual complaint about 2e style multi-classing or gestalt is that these somehow increase the system mastery gap: You risk having situations where one dude rolls a wizard/warblade while other guy comes to the table with a fighter/monk. But assuming similar levels of optimization, these options probably increase the fun, specially if the group is small.

That problem exists in the normal 3E multiclassing rules.

I don't know. Maybe this is just more complexity than is needed for what little, if anything, is gained.

[Username17]

The progressions just don't matter. It's so incredibly irrelevant whether your character has Good BAB or Medium BAB that I gave the Tome Soldier a Medium BAB just to make that point.

What matters is whether your total package results in you doing level appropriate things. Since the difference between Good BAB and Medium BAB is only +3 on attack rolls at 12th fucking level, it just doesn't matter. Everything on the level chart basically means jack shit.

The reason that 3e multiclassing doesn't work at all is that since the only thing that matters is whether your abilities add up to doing level appropriate things at whatever level you're at, so obviously an ability packaging you might take as your third level and might take as your 12th level is unlikely to work for both. Really the only way to make characters in 3e that fill hybrid concepts and are playable at several levels is by making new progressions that have level appropriate actions available to them at each of those points. Hence the Tome Soldier. And like the Tome Soldier, it doesn't fucking matter at all what kind of level progression you give these fuckers. You can make a fully effective or totally ineffective wizard or warrior or hybrid character that has good BAB, bad BAB, or in-between BAB. It's a very unimportant part of the total package.

In order to make Fighter/Wizards that are playable at level 8, you make a new class that is conceptually a Fighter and a Wizard and has level appropriate actions to take at 8th level. But whether such a character class has one BAB progression or another or elects to have good or bad Reflex saves or whatever is just not remotely important. Fucking around with alternate number progressions is just jerking off.

-Username17

[Korwin]

So, how well would this work in practice? It's simple, and relatively easy to implement. I'm curious how abusable this is. The obvious benchmark is casters. They're what make the world go round. Doing this would cost you a caster level, which means you'd be behind in your highest level of spell available half the time, and half the time you wouldn't be.

Not shure about how I think about the averaging and rounding down.
Sometimes you are hurt by it (Warblade + Swordsage), sometimes (Warblade + Crusader) you are not, if you care about BAB and saves.

The powerful thing is, you get both class features.
Which seems only worth it for spellcasting (and maybe Book of 9 Swords classes).

Also, I would not call it simple. You may not need many words to describe the rules. But building an character with it (and the different combinations) is not that simple...

[RobbyPants]

The chassis wasn’t my biggest concern. It was gaining abilities from two classes at once, one level behind.

But yeah, new classes is probably the best approach.

[Username17]

Imagine for the moment that you carefully created a complex formula for figuring out what the starting gold or the suggested starting ages were for a new class.

Whether the formula "works" or not, I don't actually care because it's dumb.

-Username17

[Iduno]

Multi-classing in 2E was also restricted. There were weird racial restrictions, because Gygax. The sane restrictions were (IIRC) that you couldn't choose a kit for your character (you're a thief, not a swashbuckler, unless you're a gnome illusionist) and you couldn't double-up on skill types (no fighter/rangers).

The result was you could be a fighter/mage, mage/thief, fighter/mage/cleric, or fighter/mage/thief. Maybe one or two others. You don't run into duskblades abusing their touch magic, because you can't be a duskblade as a multi-class. It also meant that there were few enough multi-class options that they could just write up the xp and class feature tables if they wanted to.

So I guess explain what rules you used, so people can follow the 3E "your GM may, but never will, create similar new classes that fit what you want to play", but write up the combinations that will get used.

[OgreBattle]

This thread goes a lot into multi classing: http://www.tgdmb.com/viewtopic.php?t=53 ... sc&start=0

I feel one should either go with "everybody multiclasses/subclasses" and have a system built for that, or "everyone single classes but there's 100+ classes to choose from that are functionally well tuned multi classes"

The latter is usually easier, writing up an Eldritch Knight and Mystic Ninja separately will take less back and forth checking than making sure Fighter/Sorcerer has the same balancing point as Rogue/Sorcerer

[jt]

If you have a rigid clock of at what levels characters are supposed to become able to handle certain challenges, you can make 3E-like per-level multiclassing work by using the appropriate level instead of starting each class at 1. Like you'd do three levels of Paladin then decide your next level should be Barbarian and use whatever bonuses are listed for their 4th level. This guarantees level-appropriate abilities by itself, and is compatible with whatever additional rules you want to throw in to limit combinatoric dumpster diving.

But for RobbyPants's original question, I think suggestions that require rewriting the whole system are a bit out of scope. A 2-level lagging hybrid option should be no more of a mess than what's already in the game.

[maglag]

If you have a rigid clock of at what levels characters are supposed to become able to handle certain challenges, you can make 3E-like per-level multiclassing work by using the appropriate level instead of starting each class at 1. Like you'd do three levels of Paladin then decide your next level should be Barbarian and use whatever bonuses are listed for their 4th level. This guarantees level-appropriate abilities by itself, and is compatible with whatever additional rules you want to throw in to limit combinatoric dumpster diving.

The problem with that approach is that it assumes each ability at each level is completely independent, whereas many classes have their abilities supposed to synergize and/or upgrade each other.

So for example if you have a cold mage class and get cold spells and then abilities to boost said cold spells and ignore enemy resistance/immunity to cold, somebody dipping into cold mage wouldn't be that useful.

Or your example of a barbarian, what happens if you dip barbarian mid-level and get a rage upgrade when you skipped the low level barbarian that actually grants rage?

[Iduno]

If you have a rigid clock of at what levels characters are supposed to become able to handle certain challenges, you can make 3E-like per-level multiclassing work by using the appropriate level instead of starting each class at 1. Like you'd do three levels of Paladin then decide your next level should be Barbarian and use whatever bonuses are listed for their 4th level. This guarantees level-appropriate abilities by itself, and is compatible with whatever additional rules you want to throw in to limit combinatoric dumpster diving.

The problem with that approach is that it assumes each ability at each level is completely independent, whereas many classes have their abilities supposed to synergize and/or upgrade each other.

So for example if you have a cold mage class and get cold spells and then abilities to boost said cold spells and ignore enemy resistance/immunity to cold, somebody dipping into cold mage wouldn't be that useful.

Or your example of a barbarian, what happens if you dip barbarian mid-level and get a rage upgrade when you skipped the low level barbarian that actually grants rage?

You can take any ability up to what is level appropriate? That sounds reasonable until you consider how much stupid book-keeping that would add. Paladin 3/Barbarian 1 is much easier than "Paladin 1-3/Barbarian 1 as level 3 because I wanted X bonus." I guess you could add a section to the character sheet where you write what class/level you took each level-up.

It's also possible that fixed classes and mix-and-match skillsets don't go together well. Are there any fantasy TTRPGs that went with something like a bunch of small skill trees you could choose from instead of classes?

[Korwin]

Going back to the first post. How would I do AD&D style multiclassing.


Before taking the first Level decide if you want to take this Option or not. If you take this Option:
Choose two or three classes and use the gestalt rules as written.

Changes to the gestalt rules:
After chossing those classes you cant change it, also no PRC'ing out of it.
If you took two classes, you get LA+1 and are not able to buy it off.
If you took two classes, you get LA+2 and are not able to buy it off.


Optinions?

[jt]

The problem with that approach is that it assumes each ability at each level is completely independent, whereas many classes have their abilities supposed to synergize and/or upgrade each other.

You'd have to design your classes around this from the first place. You can use tricks like letting everyone rage 1/day but that doesn't actually do something unless you picked up a class ability that gave a bonus while raging.

But yes, you could easily go too deep down the rabbit hole of making things modular and build 4E. You really aught to constrain a system like that instead of making the decision every single level. If you have to take levels 5-10 as a package (or levels 4, 8, 12, 16) then you can build compounding abilities within the package. I only suggested the per-level version because it's illustrative of what's needed to actually make multiclassing and level appropriate abilities happen in the same system.

You can take any ability up to what is level appropriate?

This undoes the thing that a system like that solves - it no longer forces players to always take level appropriate abilities. A better bandaid would be to have feats that fill in the blanks. Like prereq: any barbarian level 7+, benefit: class features from barbarian 1-3. This is only acceptable if it's filling in missing barbarian flavor though, and not providing base abilities that are required to make higher level barbarian abilities work.

It's also possible that fixed classes and mix-and-match skillsets don't go together well. Are there any fantasy TTRPGs that went with something like a bunch of small skill trees you could choose from instead of classes?

I think Fantasy Craft is supposed to work this way, but it's also such a mess that I'm not 100% sure it does.

My own homebrew replaces classes with feat chains, but making this work cleanly requires that every feat has roughly the same power level, which in turn means you have to have a diminishing returns power curve. I wanted that, but D&D really doesn't.

[Omegonthesane]

Going back to the first post. How would I do AD&D style multiclassing.


Before taking the first Level decide if you want to take this Option or not. If you take this Option:
Choose two or three classes and use the gestalt rules as written.

Changes to the gestalt rules:
After chossing those classes you cant change it, also no PRC'ing out of it.
If you took two classes, you get LA+1 and are not able to buy it off.
If you took two classes, you get LA+2 and are not able to buy it off.


Optinions?

LA is utter, total, complete, irredeemable dog shit. The game engine breaks if the party hits a Blasphemy or a Stinking Cloud while having different numbers of hit dice.

The intended positive impact of LA, that of being X levels behind for class purposes, is achieved by simply having to take a conventional level in each of your multiclass before taking the Multiclass class, or at a stretch requiring you take Commoner hit dice instead of LA.

This will still have occasional breakpoints if you find two class features that stack "properly" but all I'm coming up with is Rogue Ninjas who meet both sneak attack and "ninja thing that isn't technically sneak attack" conditions when attacking so add more d6s than they're meant to at each level - and that still turns attacks into save-or-die when casters were already on save-or-die.

[deaddmwalking]

LA is utter, total, complete, irredeemable dog shit. The game engine breaks if the party hits a Blasphemy or a Stinking Cloud while having different numbers of hit dice.

I'd offer that the design of Cloud Kill (as opposed to Stinking Cloud is itself shit. Blasphemy would generally mean that if you have an LA of +1, you would be weakened while your higher level friends are dazed. If a character with 5 REAL levels and a +1 LA is treated as a 6th level character for all purposes, you wouldn't even have that issue.

Stinking Cloud doesn't have any level-dependent effects; it's a straight fortitude save. The lower saves for a party member a level below the party might be an issue, but they're no bigger than a 'good save'/'bad save' difference. Cloud Kill automatically kills creatures with 3 HD or fewer - that's pretty grossly unfair - it should be a save to negate like Poison. If 1st level character has a +12 Fort Save and a 4th level character has a +2 Fort Save, it just doesn't make sense that the 1st level character is more likely to (ie guaranteed) to die.

Things like this come from the hodgepodge way D&D has developed over the years.

[tussock]

AD&D Holy Word would Kill 1-3 HD, Paralyse 4-7 HD, Slow 8-11 HD, and Deafen 12+ HD (for 50% spell failure). And they got that at 16th level, where it's OK.

3e Holy Word is first cast at 13th level, so Monte Cook made it CL-10 max to kill, CL-5 max to Paralyse, CL-1 max to weaken, and everyone else deafened; which is 1-3, 4-8, 9-12, and 13+ similar to AD&D, but then goes up in level because Monte Cook loves him some spell casters, and then 3.5 gives it to Balors at CL 20 and everyone dies.

It's just a bad spell design in 3e, you can't have auto-kills that scale with level. It's fine to shred tiny guys when you're up near the top levels, it takes too much out of the game to ever do it to packs of 10 HD monsters.

[OgreBattle]

There's a thread on here where it's proposed that your two components of character class are...

1) Combat
2) Out of Combat

So you may be...
1) Melee/Ranged burst/steady
2) healer, info gatherer, and other stuff