Sunday, September 17, 2017

Multiclassing, Part 2: History of the Multiclasses (3rd edition)

Last time, I diverged briefly from my discussion of D&D multiclassing through the editions to mention that my group and I gave up on D&D for a while in between the Player’s Option series (which some called 2.5e) and 3e.  To be honest, I completely undersold that.  We had whole-heartedly abandoned D&D.  It was a moribund system, weighed down by the accumulation of mismatched, overly complex subsystems over the course of 20 years.  There was no consistency anywhere: sometimes you wanted to roll high, sometimes you wanted to roll low; sometimes a higher score on your sheet was good, sometimes it was bad; thief skills and non-weapon proficiencies did pretty much the same thing but had no relation to each other whatsoever; one ability score was clearly better than all the others (Strength) and one was clearly worse (Charisma); you were never sure which dice to use for weapons; and, if you were able to figure out whether anyone was surprised when your party, which contained both a ranger and an elf, met a drow, an elven cat, and a giant eagle, at night, then you were a certifiable genius.

So we tossed out all our old books (figuratively, of course) and brushed off our hands and said “well, that’s that!” and assumed we would never play “real” D&D again.  And then ... third edition.

D&D 3e fixed a lot of problems: it positively dripped consistency, and all the abilities were useful again, and you always wanted to roll high.  But, honestly, for me, what truly drew me back into the fold, what definitively made me put on my best Pacino and proclaim that just when I thought I was out, they had pulled me back in, was the multiclassing.  It was a thing of beauty.  It was simple to understand, simple to do, you didn’t have to plan out your entire adventuring career before you ever threw the first die and you never locked yourself into dead-end paths, and you could combine almost anything with almost anything else, in a great profusion of choices.1  Not only could I be a figher-wizard if I wanted to, I could be a fighter with just a little touch of wizard (or sorcerer, perhaps), or a wizard with just a little touch of fighter, or a mostly-wizard-rogue with 2 levels of ranger, because I was an outdoorsy sneaky magic-slinger.  That “nightblade” character my fevered imagination had been trying to put together for so long was now not only possible, it was trivial.

To be clear, we’re talking about multiclassing with base classes here, not prestige classes, which are entirely different and shouldn’t be considered at all in discussions of multiclassing, in my opinion.  Prestige classes used the multiclassing rules, but discussions of whether multiclassing worked in 3e or not that focus on the eventual mess that prestige classes led to are completely missing the point.  Take prestige classes out of it entirely: saying multiclassing in 3e didn’t work because prestige classes didn’t work—whether you agree with that statement or not—is pretty much the equivalent of saying that classes in D&D don’t work because the ranger is a hot mess.2  As a simple example, prestige classes have many of the problems that multiclassing in general doesn’t: you do have to plan out your entire career, and build traps and dead ends are common.  So let’s ignore prestige classes for purposes of this discussion.

The amazing thing about 3e multiclassing—and we may as well throw in 3.5e for that matter, because the multiclassing rules were just about the only thing that didn’t change from 3.0 to 3.5—was its flexibility.  You certainly could plan out your entire 20-level career if you wanted to (such templates are usually called “builds”), but you could also just keep taking fighter levels until you got bored and then take a random level of whatever.  You could take two classes, or three, or fifteen (if your GM allowed enough splat books), and you could take them in whatever proportions you liked.  If you wanted to be half cleric and half barbarian, you could take 10 levels of one and then 10 levels of the other, or you could alternate back and forth for your entire career, or you could take 3 levels of barbarian and then 8 levels of cleric and then 4 more levels of barbarian and then two more levels of cleric and so on.  For nearly every combination of classes you can imagine, a quick Internet search will almost always get you people railing about it being completely overpowered, and also people sneering about how useless it is.  The truth is, flexibility breeds complexity, and complexity can be a good thing ... for instance, complexity is what makes it not always clear whether a certain choice is good or bad.  And that makes it an interesting choice.  Because a choice where the answer is crystal clear—where it’s always A, or always B—is a boring choice.  In many ways, it’s a non-choice.  The only reason for picking the suboptimal choice is to be different, and being different only for the sake of being different is not the best strategy.3  So, in this way, the complexity is a positive.

But complexity has a negative side, of course.  Complexity often confuses people, and leads them to making poor choices.  Let me be clear that I consider this different from being a “trap,” as that term is often used when talking about character building.  To me, a trap is a feature that seems good on paper but in practice turns out deliver way less than advertised.  That’s different—subtly different, perhaps, but distinctly different—from a case where the user has so many options that they just can’t process them all, and end up picking the wrong one, or overlooking the right one.  If it’s obvious from reading the feature that it was the right choice (or the wrong one), that’s not a trap.  It’s just that there’s so much to read that it’s easy to skim over something and not pay close enough attention to realize you’re heading down the wrong path.  Multiclassing in 3e was certainly guilty of that.  There were 11 classes in the PHB, and Wikipedia lists 42 others across 14 other books, not even counting NPC classes, core class variants, or setting-specific classes.  That’s a metric shit-ton of material, and if you were really faced with choosing one class among 50+ every level for 20 levels, that’s over 9½ million billion billion billion different possible character builds.4  Most of them are silly, sure, but the point is that it’s easy to miss things amongst that much material.

But really complexity wasn’t the problem, in the end.  Flexibility leads to complexity, sure, but the thornier issue is that flexibility leads to abuse.  I’m not going to get too deep into the actual issue of using multiclassing for powergaming—that’s a broader topic that deserves its own post (which it will get, later in the series)—but, regardless of the reality, the perception that multiclassing acquired at this stage of its history has dogged it forevermore.  The practice known as “single-dipping” (or, more rarely, “double-dipping”), which is taking one (or two) levels of a class just to get its early features, explicitly began after the release of 3.0.  It forced class designers to change the way they laid out their class features—if you put too many interesting or desireable features at low levels, people would “dip” your class, but nobody would play it straight up.  This came to be known as a “front-loaded” class, and please be sure to curl your lip in your best Billy Idol imitation when you say that, because it’s meant as a terrible insult.  3.5e fixed a lot of the worst examples of this, and Pathfinder fixed even more, but it continues to be a concern for class designers everywhere.  Now, overall I think this is a positive thing—forcing designers to spread out class features has lots of other benefits as well, such as keeping high-level play interesting—but there’s no denying that practices such as dipping gave multiclassing a bad name that it’s still trying to shake to this day.

Moving from the big picture to the smaller one, 3e multiclassing did have a number of minor problem.  Happily, Pathfinder fixed nearly all of them.  The biggest minor problem with 3e/3.5e multiclassing, in my opinion, was the XP penalty.  This was a little nose-tweak to discourage everyone from multiclassing all the time, and it could be offset somewhat via a race’s favored class, which restored a tiny bit of 2e’s racial multiclassing preferences by making an elven fighter mage progress normally, while a human one lagged behind (but only by a little).  The problem with this was that it didn’t really work.  XP penalties are always a giant pain in the ass to keep track of, so the vast majority of gaming groups just threw them out, and it was a multiclasing free-for-all.  Pathfinder wisely determined that this problem required a carrot, not a stick: in other words, don’t discourage multiclassing, but rather favor single-classing.  To that end, they added “capstones,” which are extremely cool features5 that you can only get by sticking with one class all the way to 20th level.  Of course, that only works if you actually play to 20th level, which no one ever does.  But Pathfinder foresaw that niggling problem as well, and retooled “favored classes” to be decoupled from races and just reward all characters for taking levels in their original class.

Pathfinder also removed the last few class-combination restrictions.  And the last thing they “fixed” was to make a lot more things depend on your class level as opposed to your character level.  I put “fixed” in quotes there, though, because while that curbs a lot of abuse, it also severely nerfs multiclassed characters in some situations.  The big thing that Patfhinder didn’t fix was the saving throw problem.  If you took 3 different classes that all had the same “good” save, you suddenly had a +6 at level 3, which was insane.  Contrariwise, if they all had the same “bad” save, you were stuck at +0 at 3rd level, which was just as silly, except in the other direction.  There were various schemes for fixing this,6 but really we just had to wait for 5e for it to be fixed properly.

Overall, the 3e rules for multiclassing, especially as refined by Pathfinder, continue to offer the best, most flexible roleplaying experience.  Unfortunately, “most flexible” is not always considered a compliment, and the reputation of multiclassing as “for munchkins only” is directly traceable to this version, so we can’t claim perfection.  Next time, we take a look at what went wrong in 4th edition.


1 The only classes you couldn’t combine were those with conflicting alignment restrictions, such as monk and bard, or paladin and druid.

2 And nearly always has been: note how it was annoying in 2e, broken in 3e, barely tolerable in 3.5e, weirdly off-kilter in 4e, and downright wimpy in 5e.  But I digress.

3 As opposed to being unafraid to be different for other, more valid reasons, which is often a great way to go.

4 Or about 9.5 thousand decillion, in American mathspeak.  If you happen to be British, you would say 9,500 quintilion.  I believe.

5 Theoretically.  Obviously some classes hit that target more accurately than others.

6 E.g. the “fractional saves” rule in the 3.5e version of Unearthed Arcana.