Multi-classing in level-based games

[Woot]

What’s a workable multi-classing system? DrPraetor brought it up in the OSSR: Frostburn thread but I think it deserves a thread of it’s own. Let me spitball ignorantly a bit to get started!

I’d suggest a design goal should be “Allow trading raw power for flexibility.” The problem with this design goal is that while some players will choose to do just that, the real goal for many players is to leverage the synergies between 2 (or more) classes to create a character that is more effective than the character would have been if it had been kept single-class. There’s also the question: If trading levels in class A to take levels in Class B produces a more effective class A character… is class A a busted design?

It also seems like there’s two broad types of paths such a character would pursue: either a mostly-balanced blend of skills where you’re at class A as you are at class B, such as a “sneaky backstabbing fighter” and a “few level dip” where you’re trying to add a specific limited skillset - say, a high-level rogue who wants to take a couple of levels in wizard to get a few handy illusion spells, but is still mostly a rogue. Maybe call these the 5/5 and 8/2 paths; are there others?

I’ve always felt that 3/3.5 handled the 8/2 path fairly well, but totally breaks down on the 5/5 one. Such a character rapidly becomes level-inappropriate once the party reaches the mid-levels. Perhaps due to AD&D 2e-inflicted brain damage during my youth, I’ve always thought that the “practical” effect of it’s “split XP between classes” model handled such characters more gracefully. In general, a fighter/mage would lag a level, or at worst, two, behind the pure fighters and mages in the party, which means the character still has some utility to the party, but doesn’t generally outshine the pure classes.

Operationalizing this for a 3/3.5/5e paradigm, the formula should be something like: Level in class(es) = Character level - (number of classes -1)) or perhaps more plainly, “Whenever you level, you have the option of staying your level but also being that level in a new class; i.e. a 5th level fighter could, at level time, elect to become a 5th level fighter/5th level wizard.

I’ll readily concede this provokes some really weird WTF moments from a story perspective - SUDDENLY I’M A WIZARD, HARRY! - but it “works” in the sense of letting you multi-class while remaining level appropriate. I’m also willing to concede that this may be too much of a pants-on-head crazy power-up, and it totally ignores people who just want to take the 8/2 level path (by turning it into a 9/9 path.)

[The Adventurer's Almanac]

Nice thread, I shall also spitball ignorantly to provoke discussion!

I've been playing around with the idea of a 4-tiered, 20 level system (just call me a hack now) where you choose two classes at character generation, and they both add onto a blank character. You can progress in either class as you level, but not beyond a certain point, which is determined by your tier. For example, you couldn't take more than 3 features from either class at tier 1, and more than 6 at tier 2, and so on. You finish progressing in both of these classes by the end of tier 3. If you go onto tier 4, then you have to take a 5-level long advanced class that does crazy, level-appropriate bullshit.

I'm sure this has its own issues, but I like it more than the open multiclassing madness that PTU already offers. That shit is just wild. A maxed out character will have 4 full classes' worth of shit on their character sheet... and still not be able to do much!

[deaddmwalking]

So I will concede that AD&D style multi-classing generally worked.

If you were a Fighter/Wizard or a Fighter/Wizard/Thief you would lag a level or two behind the party, but generally would be roughly as powerful - the benefits of having more flexibility didn't leave you woefully behind the curve.

But a Fighter 3/Wizard 3/Rogue 4 wasn't a '3rd level character' and definitely wasn't a 10th level character - maybe '5th level equivalent'. So that type of multi-classing comes with some baggage around determining your true 'effective level' for other purposes. That's not an unsolvable problem, but it is a problem.

In 3.x terms, adding a new class always adds a HD and associated things. But 3.x also uses increasing XP. Going from 5th level to 6th level costs more XP than going from 1st level to 2nd level. I think you could use something like that to make a workable system...

First, look at whatever class is your 'highest level'. That is your character level.

Second, whenever you gain XP, you can allocate it to your primary class or a different class. Ie, if I'm a 5th level Fighter and I gain 2000 XP, I can apply it toward going to Fighter 6th or Wizard 1.

When you add a level in a class. you don't gain a HD. You don't add all the BAB (ie, a Fighter 1/Ranger 1/Barbarian 1 has BAB +1, not +3). You basically treat each feature like gestalt - if one of your classes has a better feature you take the better. So a Fighter 2/Rogue2/Wizard 2 has 3,000 XP, but Fighter Hit Points, Fighter BAB, and all the Fighter Bonus Feats, Rogue Sneak Attack, and Wizard spells. The single class companions have 3,000 XP and are 3rd level in their primary class (meaning they have an extra HD, the next level of spells, etc).

This would not preclude you from starting a multi-class journey at 10th level - where you have 45k XP. If you spend your next 10k XP on picking up Wizard skills, you'll advance to 5th level Wizard by the time your companions hit 11th level. Advancing Wizard up to 10th might be considered prohibitive - it would take you 35k XP more (90k total) which would put your companions just shy of 14th level.

For all the levels that matter, it should work out pretty well.

As far as 'now I'm a wizard', if you're starting it with 1st level spells, rather than jumping right to 3rd, it'll feel gradual.

[Whatever]

The entire point of a level-based system is that characters get level-appropriate attacks and defenses. Open multiclassing is not compatible with that.

"trading raw power for flexibility" is a terrible design goal. If a character gives up "raw power" then they are no longer level-appropriate.

The point of multiclassing should be to allow more character concepts, especially since we don't want to write 15,000 base classes (or 15,000 Elothar, Warrior of Bladereach prestige classes). By letting players mix and match, we increase the number of distinct characters with considerably less effort on our part--to the point where it becomes plausible to sketch out most of the conceptual space within the rules as written.

[deaddmwalking]

The entire point of a level-based system is that characters get level-appropriate attacks and defenses. Open multiclassing is not compatible with that.

That's not strictly true. If every ability is level-appropriate at only 1 level and then is useless, that might be true, but that's usually not the case.

We know that a Wizard gets higher level spells faster than a Sorcerer. We know that a Wizard is better because of this. But a 3rd level spell is 'level-appropriate' at 5th and 6th level, and should see use at 8th and 9th level, too.

If you gain an ability later than other characters, it doesn't mean it isn't appropriate when you gain it, especially if it is one option among many.

[Username17]

Open Multiclassing means that you have only one class. The things you call "classes" are functionally feats.

If the abilities of Sorcerer 4 require that you take Sorcerer 3 in a previous level before you can take them, then your abilities are laid out in feat chains.
If the abilities of Sorcerer 4 do not require that you first take Sorcerer 3, then you have feats with no chains.

That's it. That's the entire mystery of Open Multiclassing explained. It's a bad idea for the same reason that hiding everyone's relevant abilities in "big feats" on big lists is a bad idea. It makes it hard to figure out how to make the character you want and easy to lose the game in character generation.

-Username17

[Foxwarrior]

I agree with what Frank said except the last two sentences. Charop and finding new interesting synergies is fun for its own sake, especially if the combo is more interesting than just massive damage, and in moderation it can even add to the fun of actually playing the game if your fellow party members have a few tricks up their sleeves unique to their builds.

[Emerald]

It's not just that level dips are similar to taking big chunky feats, open multiclassing via feats is already kinda sorta something you can do in 3e, depending on what class(es) you want to multiclass into. The later books went out of their way to let you pick up a handful of abilities from their subsystems via feats instead of multiclassing, sometimes just low-level stuff (Bind Vestige/Improved Bind Vestige for binding and Magical Training/Precocious Apprentice for arcane casting) and sometimes near-level-appropriate stuff (Shape Soulmeld/Open [Least/Lesser/Greater] Chakra/Bonus Essentia for incarnum and Martial Study/Martial Stance for initiating). Certain class features were made into feats in a limited or weaker form (Wild Cohort for animal companions, God-Touched for turn/rebuke undead, Draconic Aura for dragon shaman auras, etc.)

So if you extend and generalize that concept to more classes and allow taking closer-to-level-appropriate abilities with less of a feat investment, you can remove the need for the "8/2 path" from the OP in the majority of cases. Kind of like 4e's feat-based "multiclassing" attempt, but without the intentional nerfs and arbitrary restrictions that made that version practically unusable.

[merxa]

Ideally multiclassing would be trading depth for breath with power being roughly equivalent.

[nockermensch]

There are two kinds of multiclass character concepts.

The first concept is the character that acts as if they belong to two different classes, all the time. I think the archetypical example here is the fighter/mage, that wants to kick ass with both weapons AND spells.

For this concept in a 3E like game, simply allowing 1E style multiclass characters (what 3E calls "gestalt") seems like an easy and balanced enough choice. You don't even have to split the xp gain between classes: this should be made balanced enough by making it cost LA +1 (without LA buyout). So, a "FighterWarblade/Wizard 7" has 6HDs and the class features of a 6th level warblade and a 6th level wizard. The reasoning here is that action economy, SAD and limited character resources like feat and magic item choices should work together to limit what such a character can actually do much better than a single classed character, anyway.


The second concept is the "used to be a rogue, then became a wizard" character, in which they have a bit of early training in some class, but then change careers.

For this concept, maybe a system inspired on 1E dual class can work: The initial rogue levels are paired gestalt style to the initial wizard levels, and then from a point on the rogue side ends and it's just pure wizard. The price here is that the first level of a new class costs 1,000 exp and you have to advance one class at a time (making it a sucker's choice to keep more than one class at the highest level).

Ex: a player levels up a character to rogue 5, then decides to take the first level of wizard. 1k xp later they get to write up "| wiz 1" on their character sheet and they remain a 5rd level character until they get enough XP for the 6th level of wizard. The single-class characters at his party will be level 7 by that point, but critically, the multiclass character will only be one spell level behind and should have enough in skill ranks, class features and items to actually be a part of the adventures. And since we're not giving this kind of character a LA, they should eventually catch up with the single classed characters some time later.

The reasoning here is that low level powers are just this: Low level. Giving a wizard evasion or like 15 additional hit points (from those days as a barbarian in their youth) is kind of nice, but not campaign-shattering. It certainly should not force said wizard to be two spell levels behind forever, in any case.

[Username17]

The big problem with 1e style multiclassing and dual classing is that it's all fucking gibberish and the XP costs don't make a damn bit of sense and none of the classes have coherently defined roles. It doesn't translate especially well to later editions because for it to work you'd really want single classed role protection and multiclassed synergy to be very finely balanced at each level - which is of course not a thing.

While there are definitely people who want to play a Fighter/Wizard, it's not at all obvious how that concept is mechanically distinct from Cleric or Druid. That mostly people are just wanting to get away from some of the weird thematic baggage of the Cleric and perhaps we'd be better off if there was just a Jedi class that wasn't shit that people who wanted to play that archetype could play.

-Username17

[Dogbert]

This is a very old page, on the smackjeeves site that was abandoned (the comic moved to Comic Fury last year), but it still states everything I could say on the matter.

Basic, single-digit cure spells mean jack in games where everyone's HPs are in the triple digits and their damage in the double digits, period. You're not "versatile," you wasted a level and gimped your character.

[Ancient History]

Part of the problem is what constitutes a class and how it works. D&D classes don't generally represent distinct roles in the game - Cleric is not synonymous with healer, there's usually a fair range of variation within classes, and the class abilities themselves are often not based on overall character level.

So look at a multiclassing system like Earthdawn had it: the levels themselves rarely give you a bonus or ability, but they set the marker for the highest ability you can have. You can be a 1st circle Warrior an entire game and get your Melee talent up to legendary ranks, and you can be a 5th circle Warrior with bare minimum ranks in your talents to get you to qualify for the next Circle, and those are both legit character paths. In this case, multiclassing isn't quite feat-chain-ish, but it also avoids most of the perceived benefits of the class system in D&D3.x standards by not pretending that characters at a given Circle are supposed to be equivalent to each other in terms of power.

[Iduno]

The entire point of a level-based system is that characters get level-appropriate attacks and defenses. Open multiclassing is not compatible with that.

"trading raw power for flexibility" is a terrible design goal. If a character gives up "raw power" then they are no longer level-appropriate.

True. I guess if you want flexibility, go with a point-buy game.

I can even imagine a point-buy system where being a wizard isn't the same number of points as "hit thing a little better," but I don't think there is one. You'd have to treat a new type of spell like you would a new type of skill (invisibility replaces hiding, but not move silently).

[Username17]

Point Buy in games like GURPS does not make point costs that correspond to power level. That is, the reasoning for why something costs more points is often opaque. Sometimes things cost points for being 'powerful' but other times things cost more points for being 'hard to learn' or 'especially rare.' Long before we get to issues about mathematical cost progressions and how they interact with a random number generator, there's the simple fact that there's already not even the expectation that a character that costs 160 points be more powerful than one that costs 150 points.

Now let's talk about something that's really important: some abilities are meaningfully divisible into power levels and others are not. Shooting a firebolt could do incrementally more damage, and you could imagine facing an opponent against which a smaller firebolt is meaningless or a larger firebolt is OP. But consider something like flight. Sure, you might be able to fly faster or higher or longer or with more agility or some fucking thing, but most of what you want to do with flight is simply a check box. If you just want to get across the pit of whirling death, it legit doesn't matter if you could do it quicker or more elegantly or something.

And the level dependency of these things is a choice. Are there epic level locks? Epic level sentries to sneak past? Epic level traps to disarm? If so, why? A lot of the classic 'Thief Skills' could very plausibly simply run out of road. If you can pick the locks that actually exist, there isn't any 'better' that you can meaningfully get at lock picking.

-Username17

[Foxwarrior]

I think actually if you're going for a point buy game, it's the abilities that need point investments to scale with level that are more of a dangerous menace that needs to be abolished. Lockpicking might always be a tiny ability that establishes you as slightly more useful to have around than a random chump, but firebolts are a garbage waste of points until you hit the threshold of them being strong enough to fight the enemies you encounter. So if you, say, put "power level" and "number of points" on separate scales, like some point-based systems already do, you could then have "lockpicking" cost 1 point and "level-appropriate fire bolts" cost some larger but also non-scaling amount, like 3 or 10.

[Orion]

Non-scaling combat options depreciate really fast, but utility perks also depreciate in most games. Lockpicking and Swamp Survival might be about as good at level 6 as they were at level 1, but both are probably useless at level 11.

[hyzmarca]

Personally, I've become quite fond of classes as equipment.

If you become a level-appropriate wizard by putting on your wizard hat, then party composition can be tailored to current needs. And you can cater to the player who wants to cast spells sometimes and use a sword sometimes without having to balance a fighter/wizard.

This works best when classes have a small number of potent and non-overlapping abilities, worse when there are huge overlapping lists of abilities. But in both cases it avoids dumpster diving synergies while still allowing the player to have multiple roles.


I'm also going to say that I've come to the conclusion that power balance is meaningless in PvE games. The players are supposed to win and the outcome is ultimately meaningless either way. Difficulty is an illusion. The only question is if some players feel like they're rendered useless while the other players have the spotlight. For that, its not necessary for the players abilities to be balanced, only for them to facilitate contribution.

[Foxwarrior]

Lockpicking and Swamp Survival might be about as good at level 6 as they were at level 1, but both are probably useless at level 11.

Yeah, that is a bit of a problem. Hard to fix it in a point buy system without stacking on another kind of restriction.

party composition can be tailored to current needs

I hate this. The most meaningful decision a player can make when making a character is to help decide what the party can't easily do and thus actually regards as an interesting challenge.

[Whatever]

"level-appropriate fire bolts" cost some larger but also non-scaling amount, like 3 or 10.

How do you determine what's "level-appropriate" in a point buy system? There's no levels, and everything else is independent variables!



(You shouldn't be giving players the option to buy attacks that are too weak or too strong in the first place, but that means you can't particularly scale the opposition either)

Difficulty is an illusion.

The whole game is an illusion. Facilitating and preserving that illusion is one of the most important jobs a game designer has. This isn't just Angel Summoner and BMX Bandit (though it is that), it's also the whole idea that players are making meaningful choices. The game needs to be more than "right hallway or left hallway?" and it also needs to be more than "press win button y/n?" The rules exist to impart meaning--to the situation the players are facing and to the choice they make in response. Balance helps ensure that the questions and answers are meaningful ones.

[Foxwarrior]

How do you determine what's "level-appropriate" in a point buy system? There's no levels, and everything else is independent variables!

D&D 3 is called "level-based" even though you make a character by investing points from multitudinous pools into ability scores, skills, feats, classes, spells.

Shadowrun is called "point buy" even though you're limited more by caps on various investments, advantages that lock you out of other advantages, and so on, than by your points. D&D is the "class based" game even though Shadowrun is the one where you can't play a samurai/wizard.

These names are actually kind of bad and barely contain any meaning without nailing down more details. Since in the previous half of the sentence I suggested a point buy system with a "power level" statistic, I felt that I had nailed down that detail.

[Whatever]

Since in the previous half of the sentence I suggested a point buy system with a "power level" statistic, I felt that I had nailed down that detail.

Yeah, I see what you mean now. If you have levels, then you can scale abilities to level whether you picked up those abilities by leveling, from spending points, as a quest reward, or whatever.

If you want the kind of multiclassing people are talking about here, then you probably need to bake that level scaling into your system and then figure out how you assign abilities.

Personally, I'd rather write up a good selection of base classes (all of which cap at like, level 5) and then give people strong prestige class options to follow those up. So you'd be a Fire Mage that becomes a Death Knight, rather than some kind of Fighter/Wizard mashup the whole way.

[Username17]

Non-scaling combat options depreciate really fast, but utility perks also depreciate in most games. Lockpicking and Swamp Survival might be about as good at level 6 as they were at level 1, but both are probably useless at level 11.

That depends. Breaking and Entering abilities don't have to scale to be useful. In a game like Shadowrun there are still going to be alarms and locks whether those alarms summon mall cops or super soldiers. Even in games like D&D being immune to hostile environmental effects like heat and cold is something that becomes progressively more relevant at higher levels as more adventuring environments become inhospitable.

The 3rd and 4th addition attempts to treadmill lockpicking by giving players 8th level lockpicking and face 8th level locks is weird and bad. Lockpicking can just be mostly unchanged at higher levels. That's fine actually.

-Username17

[OgreBattle]

Open Multiclassing means that you have only one class. The things you call "classes" are functionally feats.

....

That's it. That's the entire mystery of Open Multiclassing explained. It's a bad idea for the same reason that hiding everyone's relevant abilities in "big feats" on big lists is a bad idea. It makes it hard to figure out how to make the character you want and easy to lose the game in character generation.

-Username17

That's a straightforward way to put it
---

So making open multiclassing/feat grab bag work... knowing the failure points is key so how to avoid them?

The "same game test" focuses on being able to overcome broad level appropriate challenges, so character capabilities to defeat things and not get stuck/die is important.

This article on "universal defenses" making the broadly diverse Guilty Gear characters work has been influential on my approach to game design:
http://www.sirlin.net/articles/designin ... uilty-gear

[Iduno]

Point Buy in games like GURPS does not make point costs that correspond to power level. That is, the reasoning for why something costs more points is often opaque. Sometimes things cost points for being 'powerful' but other times things cost more points for being 'hard to learn' or 'especially rare.' Long before we get to issues about mathematical cost progressions and how they interact with a random number generator, there's the simple fact that there's already not even the expectation that a character that costs 160 points be more powerful than one that costs 150 points.

Now let's talk about something that's really important: some abilities are meaningfully divisible into power levels and others are not. Shooting a firebolt could do incrementally more damage, and you could imagine facing an opponent against which a smaller firebolt is meaningless or a larger firebolt is OP. But consider something like flight. Sure, you might be able to fly faster or higher or longer or with more agility or some fucking thing, but most of what you want to do with flight is simply a check box. If you just want to get across the pit of whirling death, it legit doesn't matter if you could do it quicker or more elegantly or something.

And the level dependency of these things is a choice. Are there epic level locks? Epic level sentries to sneak past? Epic level traps to disarm? If so, why? A lot of the classic 'Thief Skills' could very plausibly simply run out of road. If you can pick the locks that actually exist, there isn't any 'better' that you can meaningfully get at lock picking.

-Username17

Sorry, I was on some painkillers yesterday, and wasn't clear.

I was trying to answer "no, you shouldn't do multiclass in a level-based system, because those are for giving people a set of classes that should all be able to meet a particular difficulty challenge at a particular time. If you want multi-class, you want a system designed to mix-and-match different types of powers." Also, skill challenges shouldn't be scaling, which means they're useless fairly quickly in a level-based game (although they still might outlast most games).