Tuesday, February 19, 2013

Class Balance

Let’s try an experiment – let’s design a simple RPG.

Characters have four abilities: Fighting, Spellcasting, Healing, and Thieving (covering all noncombat skill use). One of four numbers is assigned to each stat: 5, 3, 1, and 0. To use an ability, you must roll its number or less on 1d6.

Here’s a sample gang of adventurers.


To hit a foe in combat, Braggo must roll 5 or less on 1d6. To heal someone, he must roll 3 or less. To perform any sort of thieving (or use any other noncombat skill), he must roll a 1. He can’t use magic at all.

Every attack, regardless of whether it’s made with a sword, a spell, or your bare hands, causes 1d6 points of damage.

In combat, Braggo and Blippo will dominate, causing an average of 2.9 points of damage per round with Fighting and Spellcasting. Baldwin and Biggie average only about 60% of that at 1.75 points of damage per round. Instead of attacking, Baldwin can heal 2.9 points of damage per round. Outside combat, when the group needs to get across a chasm, negotiate with brigands, or find their way through the wilderness, Biggie takes the lead with his Thieving of 5.

I'd call that a balanced game. Everyone’s numbers are equal. The differences are in where they choose to put their strength and their weakness, and the spread from top to bottom is significant.

Now here’s a second game. It’s identical to the first in all ways but one. Instead of assigning the numbers 5, 3, 1, and 0, characters get the numbers 4, 3, 3, and 2. The gang looks like this.


Now the strongest characters in combat cause 2.3 points of damage per round with Fighting and Spellcasting and the weakest cause 1.75, a considerably narrower spread than before. Away from combat, the gap between the most skilled and least skilled characters is similarly narrow.

Both games are equally “balanced.” The rows and columns have identical sums within both games. In game 2, no one is ever left completely out of the picture, but likewise, no one shines so brightly at their specialty that they eclipse everyone else.

This is the question that gets debated endlessly in discussions over class balance. Can classes be balanced if someone must take a 0 or a 1 where someone else has a 5? Is such a broad spread inherently better or worse than limiting the range to 2-4? Game 1’s rows and columns all equal out, but can it be said to be “balanced?”

My answer is an emphatic “yes, but …”

I also say it’s the wrong question.

The right question is, which game do you prefer? Neither of them is true or false; they’re just different. As some people prefer soccer over football, some people will prefer Game A over Game B, and vice versa. If you like classes with rigid walls around them, you’ll probably prefer Game A. If you want classes that claim specialties without fencing off monopolies, you’ll drift toward B.

OD&D and AD&D are pretty clearly in the Game A mold. 4E follows the Game B mold. 3/3.5 starts out like Game B but quickly morphs into A as characters specialize through the stratosphere. Where will D&D Next land? More importantly, where would you like it to land? Which type of game do you find most satisfying?

Is it possible to have both types of characters at the same table? Absolutely ... maybe. It's actually a trickier question to answer than it appears. I'll delve into that next time.


  1. Clearly, as an AD&Dman, I enjoy type A much more. "Balance" as it is used today is a weird concept. I've been told that 2e isn't balanced because if a fighter and a wizard come to blows they aren't equally balanced in combat.

    Now, I'm much more interested in the version of "balance" which requires a balanced party in order to survive.

    Of course, this also ignores the fact that not all wizards (or fighters) are equal or even the same. Wizards know different spells, Fighters are specialized (or not) in different weapons. Everything is such a morass of variants that you can't really account for so-called balance. The best you can do is try to ensure your party wins by maximizing what you do have.

    Balance, in this case, is less important than an organic environment and rules that function organically.

    1. "Everything is such a morass of variants that you can't really account for so-called balance."

      This was precisely our attitude while we were building 2nd Ed. One 5th-level party could be far stronger or weaker than another. There were no effective controls. 4E tried to establish those controls. Some people loved having reins, some hated it (if I might belabor the obvious).

  2. I think your right about "asking the right question" and in the line of thinking, could you mix both "arrays" in the same game? Try and please both crowds?

    My guy likes Type A, but my head likes Type B.

  3. Whoops! Sorry about completely missing your thoughts on addressing this.

    I'm a dunce :(

  4. Using the above example, my preferred option would be that each character gets X number of points, 9 points in Example A or 12 points in Example B, to distribute how they see fit among the four stats.

    This way all characters are balanced against each other but the players have maximum choice in exactly what character they want to create. One player might want a specialist character that's awesome at one thing and another player might prefer a generalist who's good at everything. I see no reason why a party can't be composed of both specialists or generalists in any combination according to the preference of the players.

    I could be wrong about this last point as admittedly, I'm not putting a huge amount of thought into this concept but that's not a bad thing. I think as a whole this whole concept of play balance has been over thought to the point where we may have stripped all the fun out of the game entirely.

    As for which way D&D Next will go, if the trend continues, characters will get 5, 4, 3, 3 to create characters that seem even more powerful with even less distinction.

    1. "I think as a whole this whole concept of play balance has been over thought to the point where we may have stripped all the fun out of the game entirely."

      I wouldn't go that far. I certainly agree that this problem has been wrestled with in public forums so extensively and viciously that we're all suffering from fatigue. That's a testament to how tough a problem it is, if you actually hope to solve it, and how passionately people feel about it.

  5. I have pondered balance from another perspective. Does a class outshine others at the table with options to affect the situation. It is like jedi in star wars. The majority of folks want to be jedi because it is a "mary sue" of star wars. Not everyone, of course, as some will want to be Han Solo types, but heck, there will be people who want to be a cp3o translator.

    4e attempted to address that by giving spell like effects to all classes but it ran into trouble where it focused too much on combat, and it made all the classes feel the same.

    I am not sure what the solution is to that problem. As Syndrome says, "Once everyone is super, then no one is".

  6. The problem isn't the fighting, healing, and magic stats. It's the thieving stat. The fact of the matter for most D&D campaigns is that if you fail a combat encounter, you die. If you fail a social encounter or exploration encounter, you don't. So when you make your character good at those it doesn't really feel like you're contributing, because there's nothing on the line and thus the system feels unbalanced.

  7. Thanks for this column; it puts some issues I've been pondering in perspective.

    In my case, I think I'd prefer a mix of Model A and Model B, to address what I call the Idiot Savant problem. For example, in the RAW of D&D 3.5, perception skills aren't part of a Fighter's class skills. However, you would expect a guard -- who would probably be a type of Fighter -- to have better than average Spot and Listen scores, since they're essential to his job. To do that, though, he'd have to spend double skill points, which means he is in total LESS skilled (as measured in skill ranks) than his counterpart in, say, a mercenary group who puts all his points in Jump, Ride, Swim, etc. Does combat training degrade a Fighter's hearing and peripheral vision, so that they need to work twice as hard as a Ranger or Rogue? OD&D and AD&D have it worse, as various lights in the OSR will tell you, since the mere presence of the Thief class implies that other classes *can't* Climb, Move Silently, etc. In other words, these wizards, warriors, and servants of gods have cognitive deficits when it comes to ordinary abilities that anyone can learn with some practice.

    Every system has its flaws, of course. A common solution in the OSR is to eliminate the Thief and/or reinterpret the other classes as having base chances of performing common skills. Nevertheless, wholly skill-based systems -- notably the RuneQuest/Basic Roleplaying family -- allow for prodigies in one area who can also perform common, everyday abilities with some minimum competence. The group doesn't need a special character to scout around while everyone else cools their heels; anyone or everyone can participate just like everyone participates somewhat in combat. On the other hand, if the group *does* have an expert orator, second-story man, or spelunker a good GM will arrange situations where that expertise will become as critical as sword-swinging or spell-casting.

  8. Damn balance!

    If we are to have classes, then each class should be developed independently of the others. No designer should waste time thinking, "Is a level one cleric as good at healing as a level one fighter is at fighting?"

    Decide, in sweeping generalities, what it means to be a level one player character and then similarly define what that means for each subsequent level. Do this for every class.

    In 2E, a level 3 fighter was hardened combat veteran. A level 3 wizard was only just beginning to unravel the art of magic. A level 20 fighter was a legendary combatant, but a level 20 wizard was god-like.

    And there is nothing wrong with this.

    Further rambling: I hereby reject the idea of a "healer" class, and if we are to have a healer class, let it not be the noble cleric. The cleric is a god-charged fighter, to make any significant portion of the cleric's identity that of a walking self-refilling health potion is sacrilege. If you need a healing class, make bloody doctors, barbers, and surgeons.

    The MMO has stamped the ideology of "tank and spank" onto modern gaming giving us only three real classes - tank, damage, and heals. The MMO has also turned min/maxing from foul heresy to glorious gospel.

    "Game Balance" is a dirty, dirty term.

    1. While I agree that "game balance" is largely a chimera, the notion of designing classes independently sounds even worse. I *hate* D&D wizards precisely because at 1st level they can only cast one spell per adventure, while at 20th level they're wandering gods. Why don't students of magic find a safer source of XP? And who wants to be the guy who sits on his hands until there's a use for one of his three spells? (Obviously players can make suggestions and do simple actions, but it's still frustrating.) I'd rather have low-level wizards who can us a lot of very minor powers each adventure (a la cantrips or spell-point systems) and maybe help out with combat or other mundane things; I don't care whether they later become gods, glass howitzers, or the same with more options, since I've yet to play a campaign that lasted past 8th level.

      Granted, I do agree on the cleric -- a mishmash of abilities of which healing became his raison d'etre -- but take that away and he's little different from the Paladin. Really, if we're forced to have a class-based system, I prefer at most three classes: combat-master, magic-master, and skill-master. (True20 works a lot like this, with an ability to take levels of the other two classes.) I'd be willing to throw out one of those roles: magic and/or combat might be skills (as in RuneQuest), or the combat-master and magic-master may choose whatever skills they wish (as in Tunnels & Trolls), or the combat-master and skill-master might be the same class (as in The Fantasy Trip). I'm also leery of strict niche protection, so the skill- and magic-masters can fight somewhat, the combat- and magic-masters have a few skills, and perhaps even the combat- and skill-masters may have some practical knowledge of magic if not a few minor powers.

  9. Some people prefer football over Amercian football, you mean. ;)

  10. At the risk of over-analysing an example:

    if magic can only attack, then characters effectively have a 'Fighting' stat which is the better of their Fighting or Spellcasting. This means that a character with Fighting 5 and Spellcasting 0 is better than one with Fighting 5 and Spellcasting 3.

    If magic can attack and do other things, then this is also true, but also Magic is better than Fighting. So Spellcasting 5 Fighting 0 is better than Fighting 5 Spellcasting 0, and both are still better than Fighting 5 Spellcasting 3.

    conclusion: balance is hard.

  11. The premise is dumb because we don't know what the "Spellcasting" role does. D&D Designers have decided that it means "Absolutely everything, but maybe only a limited number of times per day, which may or may not be an issue, unless you cast the spell that negates that, too."

  12. I prefer Type B. The downside to Type A is that you can have a balanced party but only by having a large part of the group twiddle their thumbs most of the time.
    The danger with Spells and Thievery in games like D&D is that its borders are ill defined and Spells tend to leak into other areas while the players without Thievery tend to try to replace mechanics with roleplaying as it's usually the only part of the game that needs RP.