D&D 5e: The Starting Level Dilemma

I briefly alluded to several things last week specifically related to the fifth edition of Dungeons & Dragons (or D&D 5e for short).  I will attempt to expand on those in a few blog posts, of which this is the first.  These posts will mostly assume that you know (or at least care) about the rules of D&D, so you may want to give this week a pass if that’s not your thing (although I’ll try to make it understandable even if you’re a newbie).  As always, refer to the masthead for some really swell advice.

The thing that I’m interested in today is the question of what level to use for starting characters.  Now, even if you’re not much of a pen-and-paper (PnP) roleplayer, you’ll probably recognize the basic concepts of level and class from many RPG video games:¹ when you build your character, you pick a class (fighter, wizard, etc), then you start at level 1.  As you play, you gain more and more levels in your chosen class.

Well, assuming a single-class system, of course, which is how most video game RPGs work.  That is, once you pick your class, you’re stuck with that class forever.  Or until you change to a different character, at least.  In the first versions of D&D, that was how it worked as well.  Then they added something they called multiclassing (but that was an awfully generous term for it), and something called dual-classing, which was even less so.  Neither one was much use if your goal was to be able to trade power and specialization for versatility and generalization.

By the time D&D hit third edition, someone on the design team had cottoned on to the idea that making multiclassing stupidly difficult was annoying everybody.  Unfortunately, their solution was to make it stupidly easy instead.  You would think this wouldn’t be a bad thing.  And, indeed: when I first read the multiclassing rules in 3e, I was over the moon.  I thought that was amazing: very simple mechanically, very flexible, and it radically increased your chances of being able to build the character you wanted, no matter how wacky a combinaton it was.  But the problem is, when you make things too easy, they’re easy to abuse.  And the ease of multiclassing in 3e led to a practice called “single-dipping” (or, occasionally, “double-dipping”), where you build a character who has one (or two) levels in a bunch of different classes, just to get the cool features that those classes get at lower levels.  This became a favorite technique of “min-maxers,” or, as we like to call them, “munchkins.”²  Which in turn gave multiclassing a bad name, which in turn led to a lot of backlash.

Take 4e.³  In 4e, it was actually impossible to multiclass at all.  Oh, sure: you could simulate it somewhat, vaguely and crappily.  But not really.  This was one of the big reasons—probably the biggest, honestly—that I never actually played 4e.  It was such a massive step backwards to the bad-old-days that I wasn’t even interested in trying it.  You see, for me, D&D has always been about building interesting characters.  I like the weird classes, the unusual classes.  Thus, I’ve played many a bard, many a druid, and one monk, who was probably the longest-running character in my decades-long RPG career.  I’ve only ever played a fighter twice, and hated it both times, and I’m pretty sure I’ve never played a wizard at all.  I like options, the wackier the better, and I like creating convoluted backstories, and I like making characters who leave an impression.  For years, the mechanics of D&D frustrated me when I was trying to push the boundaries of character creation, and finally 3e came along and fixed it, and then everyone decided that multiclassing was for munchkins only and I got screwed.  Many GMs in online forums will tell you they don’t allow multiclassing in their games at all, and while of course I would never play with anyone so close-minded, here comes 4e and the option has been removed completely.  No wonder I went over so enthusiastically to Pathfinder.

Pathfinder’s “answer” to multiclassing, by the way, was to stop trying to discourage people from mulitclassing with penalties and whatnot, and start trying to encourage people to single-class, using bonuses and a “capstone” feature at 20th level.  (For the uninitiated, 20th level is—usually—the highest level a D&D character can achieve.  So, if you multiclass at all, you can never get to 20th level in any class, because your 20th level muticlassed character has, by definition, more than one class.  The problem with this technique as a carrot is that it’s beyond rare for most people to play one character long enough to hit 20th level.  Unless you don’t start at level 1, which gives you a tiny peek at how the topic of multiclassing relates to the starting level dilemma.)  I like this better, obviously, but it doesn’t really do anything to alleviate the concerns with multiclassing: it’s still a tool that can be easily abused by powergamers.

So now we get to the meat of how the whole multiclassing debate has influenced the starting level level for characters in new games I run.  Because I’m running 5e now, and 5e has taken yet another approach to multiclassing.  It’s a two-pronged approach, and the first prong consists of finally figuring out a middle ground for multiclassing.  Here’s the basic problems with 3e-style multiclassing and how they’re addressed in 5e:

• Classes often get very cool/useful features at early levels.  This is called “front-loading” a class’ features, or sometimes just a “front-loaded” class.  Front-loading seriously encourages single- or double-dipping, because you can take one or two levels and get “the good stuff” and then move on to a different class.  This is the most obvious problem with multiclassing—so obvious that D&D began addressing it in 3.5e,⁴ and Pathfinder had already made major strides toward correcting it as well.  5e just took it even further.
• Saving throws get out of control very quickly.  In 3e, there weren’t but 3 saving throws in the first place (down from 5 in 2e), and nearly every class got 1 “good” save and 2 “bad” saves.  So you only had to multiclass twice (at most) to get all 3 good saves, and the fact that everyone’s saves at first level were a bit high to start off (in order to keep you from dying instantaneously the first time you had to make a save), taking 1 level in every class meant you’d end up with ridiculous saves.  5e fixed this by expanding yet simplifying the whole saving throw system: there are now 6 saves, but they correspond exactly to the 6 ability scores, and, instead of having a bonus that goes up every level (and is different for different classes), you’re either proficient in the save or you’re not.  How this impacts multiclassing is that you only get saving throw proficiciencies as a 1st-level character.  If you take the first level of a different class at some later level, you just don’t get any additional saving throw proficiencies.  And, because you’re either proficient or not, you don’t have to worry about stacking anyway.
• Spellcasters get hosed by multiclassing.  In 3e multiclassing, you keep your spells from different classes entirely separate.  So while an 8th-level wizard is casting 4th-level spells, an 8th-level character with 4 levels of wizard and 4 levels of cleric⁵ can only cast 2nd-level wizard spells, and 2nd-level cleric spells.  That’s a huge disadvantage to multiclassing as a spellcaster.  5e fixes this by allowing spell progression to “stack” between classes, sort of.  So if you’re a wizard 4/cleric 4 in 5e, you only know 2nd-level wizard spells and 2nd-level cleric spells, but you can cast them as 4th-level spells, which increases their power so you’re not quite as useless as you would be in 3e.
• The argument against multiclassing is too weak (or so ridiculous/complex that no one will enforce it).  In 2e, only certain races could multiclass (or dual-class, for that matter).  That was so silly that most people ignored it.  In 3e, you get a complicated reduction in XP⁶ for every extra class you take.  That one is so math-intensive that most people ignored it.  In Pathfinder, they’ve replaced the stick with the carrot, but the every-level bonuses (1 extra hit point or skill rank for sticking to your original class) are trivial enough to ignore, and the capstone, as I said, is mostly a theoretical carrot, as opposed to something you could actually eat.  In 5e, they fixed this by tying ability score increases (and feats, if you use them) to your class level instead of your character level.  Which means that multiclassing forces you to delay your ability score increases, which can be a serious but simple to enforce downside.⁷
• The argument for multiclassing is too strong.  Basically, there just isn’t enough variation in the existing classes to allow many character concepts to be built.  So multiclassing is often the only way to play that sneaky, semi-mystical shadowcaster,⁸ or that ever-elusive “gish” (the prototypical fighter/mage).  Multiclassing often wasn’t perfect for these characters either, but it was a damn sight better than the alternative, which was nothing.  Strangely enough, 3e also led to the explosion of “non-core” classes (both official and third-party, not to mention homebrew), so multiclassing obviously wasn’t the best way in many people’s eyes.  But, in the core rules, it was what you had, so you dealt with it.  Pathfiner eventually introduced “hybrid” classes (still non-core) in an attempt to alleviate this somewhat.⁹  5e deals with this issue using their second prong—subclasses—which we’ll delve into in more detail in just a moment.

So multiclassing in 5e is better, for the most part.  But it is harder, which means that it can be more difficult to make that perfect character.  I’ve said (a few times, to a few different people now) that, while I’d much rather play 5e, I’d still much rather build a character in Pathfinder.  I’m not a munchkin, nor a powergamer, nor a min-maxer.  I just love having loads of options.  Anything that increases my options gets a thumbs-up, and of course anything that decreases them gets the thumbs-down.

But let’s talk about that second prong for a bit.  “Subclasses” (also called archetypes, paths, or any number of different terms) are actually quite a brilliant concept.  There’s always been a bit of a chance for specialization in certain classes: many versions of the wizard could choose to specialize in a “school of magic” (such as illusion or divination), and many versions of the cleric had to choose a “domain” (or two) representing the particular areas of interest of their chosen god.  But Pathfinder took that to a whole new level, letting barbarians choose different rage powers, allowing wizards to either have a familiar, or a special wizards’ staff, giving rogues a set of “talents” that they could choose from at various levels, etc.  Suddenly every class had lots of little choices to make, and the end result was that it was now much more unlikely that two characters of the same class would end up being just copies of one another, making at least one redundant.  5e doubled down on this, giving the character one big choice instead of lots of smaller ones.  That choice is the subclass.  So, no longer are you “just” a fighter: you’re either a champion, or a battle master, or an eldritch knight.  The three subclasses play quite differently, yet they all use the basic chassis of a fighter.  This makes adding new “classes” much easier, and much easier to balance.  You merely add a new subclass instead, giving you a much smaller set of features to delineate, and the majority of the class features are already balanced by virtue of being core, thoroughly playtested classes.  From the perspective of multiclassing, it simply removes a lot of the need.  The eldritch knight neatly fills that desire for a gish, and my shadowcaster can either be a rogue archetype (for which there already is one: the arcane trickster), or a wizard tradition (such as a school of shadow; there’s no official version of this, but a quick Internet search will turn up several potential options¹⁰).

The solution isn’t perfect though.  If we imagine that 5e’s eldritch knight is analogous to Pathfinder’s magus, and perhaps that the battle master is a bit like 4e’s warlord (and, yes, I know: neither analogy is without its problems), then you could imagine, at least if the magus and warlock were in the same system, that, with multiclassing, you could play a sort of warmage that was a combination of the two classes.  Maybe two levels of magus followed by a level of warlord, then repeat as needed.  But you can’t do that sort of thing in 5e, even with its disadvantageous multiclassing, because they’re both fighters, and you can’t multiclass with one class.  This is not the real problem though.

The real problem is that you don’t actually choose your fighter subclass until 3rd level.  Now, not all classes are like this: two (cleric and sorcerer) have you choose at 1st level, two (druid and wizard) have you choose at 2nd level, and one (warlock) is totally weird and has you choose one thing at 1st and another at 3rd.  But the majority (7 out of 12¹¹) don’t have you make a very significant choice about what sort of role your character is going to fulfill—in some cases, even down to whether you can use magic or not¹²—until 3rd level.  That’s very late, in my eyes.  That means that, for 10 out of 12 classes, a first level character of that class isn’t yet who they will be, eventually.  They haven’t yet fleshed out their concept.  And I understand why 5e does this—it helps eliminate dipping by de-front-loading the classes—and I’m not even saying I think it’s a bad idea ... but it has a pretty big downside.

And, because of that, I’ve decided that it makes a lot of sense to start 5e characters at 3rd level.  Looking back on it, I sort of hated playing 1st-level characters anyway.  Nearly everything you try to do you nearly kills you (and often actually kills you), so it’s just a stressful survival slog to 2nd or 3rd anyway.  I think most of us just started at 1st level because that’s what you were “supposed to” do.¹³  But it’s not required.  And maybe it’s even a good thing.  I’ve only had the opportunity to carry out this plan once so far—my current game with my two boys—but I really like how it worked out.  The characters didn’t have the chance to play through those early levels, true, but now they’re coming into the story with a richer history and more options.  It feels pretty right so far.

__________

1 Who in turn derived said concepts from PnP RPGs in the first place.

2 Wikipedia has more info on what constitutes munchkinism, as well as a few guesses at to the origin of the term.  But, again: if you’re already a PnP RPGer, you know the term, and, even if you’ve only played video RPGs, you’ve likely run across it before.

3 Please.

4 The “minor” update to third edition that was released between 3e and 4e, if you’re not familiar.

5 This combo is usually called a “mystic theurge,” and 3e offers a “prestige class” which helps address part of this problem.  But it doesn’t fix it completely.

6 Experience points.

7 Alternately, you can just only multiclass in groups of 4 levels per class.  But that has its own issues.

8 Attempting to combine wizard and thief/rogue was the focus for many of my homebrew classes, as well as two of my major PCs (both using the 2e Skills and Powers rules, which many call “2.5e”).

9 I’m not even going to mention 3e’s “gestalt” classes, which were so over-the-top as to be difficult to take seriously.

10 Although, honestly, I think I might be better off trying to realize this concept via warlock instead.  A warlock with any of the three core patrons—depending on what sort of flavor you’re going for—and maybe a homebrew pact boon (I just don’t feel like blade, chain, or tome works here ... not sure what the right answer is though) would probably end up feeling more right than anything I could do with either arcane trickster or a school of shadow.  But that’s starting to drift pretty far afield of the original topic.

11 Eight out of 12 if you include warlock, but that’s a tough call.  Your patron (1st-level choice) gives you more features to differentiate, numerically, but your pact (3rd-level choice) probably makes the most difference in your playstyle.  So it’s a bit of a toss-up.

12 If you’re a particularly big D&D nerd, you will be tempted to bring up 3e’s (and Pathfinder’s, for that matter) “half-casters,” the paladin and ranger, here.  But I’m ignoring those, for a very good reason: they’re stupid.  The prototypical ranger is Aragorn, and the prototypical paladin is Lancelot; did either of them cast spells?  No, of course not.  I say adding spellcasting—especially delayed spellcasting—to paladins and rangers was lazy design, plain and simple, and I’ll stand by that statement.

13 Unless you were playing Dark Sun.  Theoretically, everyone’s Dark Sun character started at 2nd level because it was just too damn hard for 1st level characters.  But I’ve often wondered if the Dark Sun designers were just tired of killing off fragile first level characters, mostly by accident.