More on Classed vs. Unclassed Games

I’ve read a lot of really great posts as a result of my hornets-nest stirring post about how classes are better than unclassed games. Some rebuttals have been pretty smart and have changed my mind about a number of things. For instance, it seems obvious in retrospect that EVE has classes — for my definition of that word — because the ship you’re in tightly constrains the role you can perform.

Other rebuttals have been frustrating, either because they misunderstood my point or because I had already specifically talked about their counter-argument already and they ignored it.

But mostly, I think problems arose from my use of the word “class”, which was not defined well enough. I’ve defined it before on the blog, but I didn’t define it in the last post. (And I attempted to tackle a bunch of small side-topics which made things murkier.) Let me focus just on the core point for a second.

What a “Class” Really Is

By my definition, a game has “classes” if the game’s roles are restricted so that you can only do one of them at a time. What are roles? Well that will differ from game to game, but most games (especially ones that want group-based static content) will have roles designed into them from the get-go.

In a classic “trinity” game, the roles are tank, healer, and DPS. So a trinity-combat game has classes if it’s impossible to perform both the “tank role” and the “DPS role”, or the “DPS role and the Healing role” at the same time.

Now, all the other verbs, the non-role-defining verbs, should be sprinkled liberally everywhere, using any structure that seems fun. That includes the crowd-control verbs and the buffing and the debuffing and the fast travel and so on. You can use a point-buy system, or a skill tree, or whatever seems like it will capture players’ imaginations the most. (You can even use lesser-powered versions of the class-defining verbs, like Aion does and like Rift seems to be doing. If they’re timer-limited or otherwise weakened, they don’t end up doing much harm.) As long as the key role verbs are restricted, it’s still classes by my definition.

A game is “classless” if you can fully play multiple roles at once, or if there aren’t any predefined roles anyway.

The Heart of My Argument

Given the above definitions, the no-drama way to restate my last post is: if you have classes you will have a much easier time keeping players within a reasonable continuum of power. That means that the gimpiest character build is not orders of magnitude weaker than the most powerful. (They may be twice as good, even three times better, sure. You’ll get that in any system. But that’s not ruinous, it’s just embarrassing.)

Keeping people in the right ballpark of power is crucial so that you can make content for them. Content is quests, items, and other things that they find fun and rewarding.

That’s it, really… that’s my argument. Debating how you get your skills (trees versus lists versus random or whatever else) isn’t really part of my argument. That’s all frilly extra stuff that doesn’t affect balance too much.

The reason I like my definition of classes is that it’s Boolean: a game either has enforced roles or it doesn’t. The Boolean nature is why I can make a broad claim. I don’t have to waffle much because there are very few cases where you want to intentionally weaken your roles.


Let me quickly go over the various counterpoints and rebuttals from the blogosphere. These aren’t all aimed at the main point… some of them are discussing side-points from the first post… and some are NEW side arguments that had good points.

What about hybrid classes?

Even WoW tends to blur the lines with classes like the Druid which can perform multiple roles at once, or at least in rapid succession. Doesn’t that counter my argument?

Well, those hybrid classes are always a nightmare to balance. For instance, the original design of the druid seemed to be that “the Druid is 80% as good as a regular tank and 80% as good as a regular healer”, but when he can’t get into groups, this design breaks down. The flip side is the tank who is also a bad-ass healer, which is a problem too. In WoW this has been a source of constant balance tension.

These classes are hard balance problems, but not as hard as they could be because they’re prepackaged deals: The designer can tweak the healer parts of the Druid without tweaking the healer abilities of Priests. So if they can figure out what the right power level is (not easy), they at least have the ability to fix it easily.

You talk about EVE, but EVE has classes!

I’ll give you that one — using my above definition, EVE does have classes because the ships define the role you can perform. If I were rewriting the blog post, I would substitute Champions Online for the examples, because it definitely does not have classes by my definition. It’s also a better poster-child for the sort of balance implosion I meant to highlight.

But What About Diablo 2!?

Some PvE games can be balanced while seeming to defy my class definition… but most are are trick examples. Diablo 2 is such a game. This game only has one class by my definition: mass-murderer. Everyone is a PvE mass-murderer.

In Diablo 2, mass-murderers can group together in pretty much any combination and be successful, or they can solo and be successful. There’s pretty much no rock/paper/scissors mechanics going on; there’s very little role-requirements (some classes have modest buffs that are better for certain other characters, but they’re so unimportant that it seems weird to even call that out).

D2 is just a collection of bad-asses soloing together. It’s fun, and it’s a reasonable way to make a game, but Diablo 2 is not an example that disproves my point, either.

Just Look at <Pen and Paper Game X>, They Show How To Do It!

Tabletop RPGs are different. When I’m DMing a game and see somebody who’s not having fun, I will pull them aside after the game and say, “I can tell you’re frustrated. Why don’t we change X and Y about your character?”

In an MMO, there’s no personalized service. The game does exactly what you tell it to, even if what you tell it to do is alienate your users and make them quit. That’s why I’m always hesitant to consider a tabletop game system as having magic bullets to MMO problems.

They are good sources of inspiration, though. I have a pretty big collection, and I’ve ordered two new ones based on blog comments.

But I like playing craptastic builds!

Are you sure? I think that you like feeling unique, and you like having lots of options. And you’re willing to play crappy builds in order to get those feelings. That’s not the same thing.

Designers can give you feelings of uniqueness and choice without leaving you leveling ten times slower than your peers.

(And yes, I mean “ten times” slower. When Champions first launched, the best builds were at least an order of magnitude (10x) better than the worst classes. When AC1 launched, the best were at least two orders of magnitude (100x) stronger than the weakest… possibly more.)

No, I REALLY DO Like Playing Gimpy Characters!

I can accept that there’s a few people who really do like playing gimpy templates. But for every person who will play a gimpy class in quiet, there are lots more that will get on the message boards and whine forever. They had a character design in mind, and the game said “play what YOU imagine!!!”, and sure they could make what they imagined, but it turns out the game fails to let them enjoy it.

They will blame you. A few will literally come to your place of business to yell at you; some will threaten to poison your dog; they can never be satisfied; they are emotional and they feel betrayed.

I’m making a very tangential argument here, but it’s one worth considering. Do you really want to let people believe they can play a certain way and then not really support it? That’s basically lying, and in my experience, MMO players as a group don’t take it well.

Magic: The Gathering intentionally creates useless verb cards because it’s fun for players to figure out the best builds. They don’t have classes, and it still works, because players like figuring out what works and what doesn’t!

Yes, very true. But as I mentioned when I pointed this out in the first post, M:tG works great because their entire product is a new batch of verbs every year. Magic doesn’t have any content that needs to stay in continuous use for the next five years.

I totally agree that letting players make bad choices in an atmosphere of complete control is a fun mechanic for many gamers, and that some level of character control should be in any modern MMO.

I also agree that any game where you have tons of free-form skill choices needs a way to undo those choices cheaply (as is the case with Magic, where you can trivially change your deck). Otherwise people will leave.

But at the end of the day, Magic’s system is not an MMO-usable system. Think about it: every year or so you basically throw out all your cards and start over. Imagine an MMO that made you redo everything every year. It’d be called ‘A Tale in the Desert’ and it might have a cult following, but would not be able to find the audience you probably imagine it would.

MMO players, as a rule, become attached to their character and their stuff and they aren’t happy about losing it. When you take their stuff away they tend to throw around words like “I want all my money back for the past year, now that you stole what I earned!” And then they call the Better Business Bureau and then you have to try to explain to them how an MMO works. It’s fun in much the same way that being trepanned is fun.

Games can’t be perfectly balanced anyway, so stop wasting time on that!

OH MY GOD please do not use that argument against me. It is my pet argument and I’ve spent half the blog saying that very thing. I think that “fun” is way more important than “balance”, and I think designers should implicitly accept that an MMO can’t be balanced.

But I made that point IN THE POST, and explained why it’s different here. So why would you arghhhhgdfgfdg.

It’s not about balancing for “fairness”. It’s about balancing for “ability to make content and keep the game running.”

In Champions, the balance was only maybe 10x off, and it was still pretty killer. In Asheron’s Call 1, the difference was quite a lot more. A well-built level 20 character could kill an army of level 60 opponents, maybe ten or fifteen of them. At the same time.

These are power differentials that goes beyond “balance” and into “how in the hell are we gonna make content for all these different power levels?” It’s not about a fair playing field. It’s about making content.

In a classless system where roles are freely mixed, you tend to get very big power differences and it tends to be extra hard to fix them. That’s because none of the roles are especially broken… but when the roles are combined, they go crazy and you can’t keep them in check. So you have to shunt in weird special-case stuff, over and over, hoping that it helps enough. It sucks and it’s very time-consuming.

You can fix it. It can be done. And doing it is a great design challenge that feels good. To repeat, you can make this sort of game work. You will, however, take a long long time doing it.

If you just jot down on your Design Document that your MMO will have a completely free-form skill system, and you don’t allocate a whole lot of extra balancing time, you’re in for a surprise! And it’s not surprise cupcakes, it’s surprise pain.

But <insert game I like> had no classes and it’s balanced!

First, ask how long it took to balance the game. Did lots of players leave before the balance was achieved? If so, that was pretty much my argument. It’s not impossible to do, just much harder to curb because you often don’t have the granularity to tune things properly.

The other common argument is in PvP games where you just assume everybody’s gonna reroll a bunch anyway. The PvP system of “rock-paper-scissors-lizard-spock, choose two” is the classic example. That “works” because most of the combinations end up being completely worthless, but who cares, PvPers will reroll anyway, and there’s very little other content that gets in the way.

PvP players seem more willing to reroll anyway. But if you’re trying to provide a bunch of PvE content as well, you’re in for a lot of pain.

Conclusion: Thanks for the discussion!

These have been the best discussions I’ve seen in a while, and I’m happy for that. I actually wrote that post long ago and then mothballed it, because… why do I need the headache of a lot of arguments? I wasn’t trolling: I stand by my premise. But I knew it would piss people off.

But I’m glad I posted it; I learned some things, and refined some of my ideas based on others’ rebuttals, and I now have like three new blog topics to talk about too.

I wish the blogosphere was more willing to pick up other, less controversial elements of game design, because defending your point of view is a great way to find out what you really believe. Unfortunately there are only a few topics that seem to get “bites”, and I tend to have to state it controversially in order to get people to overcome their apathy and reply. This is frustrating. But I’ll take what I can get!

This entry was posted in Design. Bookmark the permalink.

26 Responses to More on Classed vs. Unclassed Games

  1. Anthony says:

    Class-based systems are about choice, Skill-based systems are about calculation. Choices are better. Choice and Conflict

  2. First, definitely kudos for sparking a great design discussion on the various blogs. It’s been a while since someone posted something up so provocative that got people discussing a topic. Definitely a good thing!

    Eric posted:

    But at the end of the day, Magic’s system is not an MMO-usable system. Think about it: every year or so you basically throw out all your cards and start over. Imagine an MMO that made you redo everything every year. It’d be called ‘A Tale in the Desert’ and it might have a cult following, but would not be able to find the audience you probably imagine it would.

    Or, you could extend the cycle to once every 2 years and it’d be called World of Warcraft and be the game everyone wants to emulate. In “Cataclysm” there are armor pieces that have more stamina on them than than whole dungeon sets did back in the day. People are regularly discarding their epic purple equipment for random quest green items. Hell, the talent revamp means that your old lovingly-crafted build might not be even possible anymore. But, people seem to be generally enjoying the game just the same.

    Ultimately, I fear your position boils down to maintaining the status quo. Let’s assume that a skill-based game is massively harder to balance, and that balance is a vital part of an MMO. I’m not sure your advice is really useful, though. For a novice, sometimes allowing them to “do the impossible” will allow them to achieve things us veteran designers would never try, yet still be astounded when someone gets it to work. For the experienced designers, why should we shy away from something because its hard? Shouldn’t we stretch our abilities and expand possibilities?

    If we really wanted to be radical, we could analyze the assumption that games must have a “power continuum” and trying bandage that issue without looking at the deeper problems. Maybe taking a look at replacing this assumption might be a lot more interesting than advocating classes for all MMOs because they’re easier to design.

    Then again, I’ve always been a bit touched. The first part of my pseudonym suits me, I guess. ;) But, still, an interesting discussion.

  3. Ben says:

    Perhaps the “next big thing” is actually to design class roles, but break the “trinity”? After working and playing for a while, I’ve come to the same realization – that free-form classless characters are very difficult to balance and create content for. However, I’ve also noticed a good number of folks who are tired of the same roles in each game they play. Do you think that it’s reasonable to try to overhaul the “trinity” with new roles, or is it pretty much here to stay?

  4. Ben says:

    Or, you could extend the cycle to once every 2 years and it’d be called World of Warcraft and be the game everyone wants to emulate. In “Cataclysm” there are armor pieces that have more stamina on them than than whole dungeon sets did back in the day. People are regularly discarding their epic purple equipment for random quest green items. Hell, the talent revamp means that your old lovingly-crafted build might not be even possible anymore. But, people seem to be generally enjoying the game just the same.

    This isn’t even remotely what Eric was talking about though. It’s not like warriors somehow don’t play like warriors anymore, or that you suddenly don’t have certain roles anymore. The verbs Eric was talking about that players can do are mostly still there despite the expansion pack revamps. What Eric is talking about is throwing out major mechanics every 2 years (imagine throwing out tanking after 2 years and adding in mezzing), rather than simply changing the gear and making adjustments to class abilities.

    The equivalent to what MtG does every two years is SW:Galaxies and their NGE. That’s just not feasible, nor reasonable in this genre.

  5. Mike Grem says:

    @Ben: I don’t think Magic’s standard format rotation is as drastic as the NGE. All of Magic’s verbs pretty much remain the same, except flavored with some new mechanic or other for its block. Blue still draws well, black still has self-sacrificial power, et cetera. Magic has had some NGE-like changes (the stack’s creation, and Planewalkers) but it remains for the most part the same game. They have a stable of useful, major mechanics that stick around (I’m referring to the color pie– Mark Rosewater writes a lot about it at, if you want to check those out), with minor variations from block to block.

    In this way, I think Brian’s comparison is pretty apt. The “blue-white control” deck archetype has been around as long as I’ve been playing, and while the specific strategy used to accomplish its endgame changes from set to set, by and large its general strategy is “stall with permission until you exhaust them and win.” So, warriors still play like warriors, etc.

  6. wufiavelli says:

    You can also make classes designed for different gameplay. AOC bear shaman worked great for me because it was basically Whack a Mole to stay alive. Kind of like giving everyone their own little minigame as they grind monsters or quests.

  7. William Newman says:

    Anthony wrote “Class-based systems are about choice, Skill-based systems are about calculation. Choices are better.”

    In a sufficiently interesting system of little tactical decisions, broad strategic choices tend to emerge. Consider how RTS games support thematic choices like “to zerg or not to zerg” that are composed of lots of little tactical components. And I might be able to fill several DVDs with the books and articles written on the thematic choices supported by the tactics of Chess and Go.

    Trying to get fine control of balance in a game where the big themes are emergent behavior is probably fundamentally more difficult than getting the same control of balance in a game of a modest number of explicitly designed multiple choices. Trying to get other desirable properties, like a high probability of suitably dramatic outcomes, is also probably fundamentally more difficult. Maybe those difficulties are so great that it’s generally wise for a game designer to avoid fooling around with designs where big choices are emergent behavior, and just support a modest number of predetermined big choices imposed from above. But practicality aside, I don’t think emergent choices make for a bad game: when it is practical to design a game where big theme choices are emergent properties of the smaller properties of the system, the resulting game can be pretty neat. There exists a population of gamers who enjoy crafting their zerg strategy out of all the little tactics involved, and getting in arguments about which variant of the strategy is closest to optimum under the rules of the game universe. Not all of them are likely to be satisfied by a game where instead of zerging emerging from the underlying rules of the universe, a new rule of the universe is imposed mandating a multiple choice between a preconfigured zerg strategy and fourteen other preconfigured strategies chosen and supported by the game designers.

  8. Noah says:

    Being an avid MtG-er, I would just like to point out–in the interest of correctness–that “Magic doesn’t have any content that needs to stay in continuous use for the next five years.” just isn’t true. Legacy and Vintage formats (which are still widely played and there are regular WotC sponsored tournaments for them) are exactly that: cards that have to hold up for the entirety of magic’s future. At least in Legacy some cards can be banned to allow for balancing; Vintage has very rare occurrences of banned cards–mostly cards that don’t fit into the playstyle anymore (ante cards).

  9. Tesh says:

    On MtG, they really don’t introduce a lot of verbs each year. Maybe a half dozen, and that’s as others are phasing out from two years ago. Standard play is two years of verbs on top of the core game. There’s plenty of variety, but most of it is tinkering with nouns and letting the verbs play off of each other. It’s the interaction, not the verb count, that makes MtG tick. (Though as Noah notes, Legacy and Vintage have even more going on, something that the designers are all too aware of, and that they stay on top of.)

    As for cost of production, if you’re running a subscription game, that’s not an unreasonable pace. It’s just that Wizards has a different design focus; the game is skill based PvP, not level-loot PvE grinding with token PvP. If you had an MMO without levels or gear, you had better be producing at least that amount of variety to justify a subscription. (Or selling content packs that introduce new verbs. Arguably, this is what Guild Wars did.)

    If you want to make a WoW-like MMO, by all means, stick to the status quo. That’s not the only way to make MMOs, though, and it’s not even the best way, depending on what you’re trying to accomplish.

  10. Cat says:

    Eric said: “I think that “fun” is way more important than “balance”, and I think designers should implicitly accept that an MMO can’t be balanced.”

    Honestly, there should be more applause for this comment.

    WoW just completely redid my preferred class (Paladin), I guess in a ridiculous attempt to “balance” it. You see, Paladins were easy to play in Wrath of the Lich King. That’s what made playing one so fun for me. (They also had a reputation of being boring if what you were after was lots of buttons to push and lots of choices to make. Bored Paladins eventually rolled Warriors.)

    Now I have an un-fun Paladin with twice as many buttons to push, a second resource to manage (holy power), procs to watch for, and tons of “options”. My spell rotation went from completely predictable to unpredictable, forcing me to pay more attention to my cooldowns than the damned game.

    It seems like a bunch of hard-core designers grabbed the reins and tried to “fix” things that weren’t broken. One of our core raid healers, also a Paladin, thinks the changes were totally un-fun, too. I wonder who is enjoying them?

  11. wufiavelli says:

    Anyone ever think of making a class total based on skills that change every month?

    So anyone playing this class basically has to test out all the skills and make a new build every month.

    Good way to pull in a fringe crowd.

    I really think it would be interesting and possible to make a Game based of skills.

    Not a traditional MMO with a skill system added on. But a game total based arolund the factor of arranging skills. They will change, and people will except this because that is the purpose of the game.

  12. Pingback: Classed vs. Unclassed MMO’s « Kitty Kitty Boom Boom

  13. Tinman_au says:

    “In Champions, the balance was only maybe 10x off, and it was still pretty killer. In Asheron’s Call 1, the difference was quite a lot more. A well-built level 20 character could kill an army of level 60 opponents, maybe ten or fifteen of them. At the same time.”

    In a robe…you forgot to add “in a robe!” :o)

    Of course, that was one of the things that made AC such a great game too, wading in and killing dozens of monkey’s/Olthoi at a time…it’s truly one of the only MMO’s I’ve ever played where you could feel your toon was heroic…

  14. Eric says:

    @Tesh – we just have a terminology mismatch. I would say every usable skill in an MMO is a “verb” because the player chooses to do it. But in the context of a card game that’s awkward phrasing and I think you would instead clump e.g. most of the skills that your typical DPS class into a couple of “verbs”.

    So I’m just saying that M:tG makes hundreds of usable actions every expansion, and they spend quite a lot of time picking them and playbalancing them.

    To use your form of the word “verb”, I’d say MMOs very rarely create new verbs in that sense. An MMO expansion that introduced a half-dozen new kinds of actions? That would be pretty awesome, but pretty rare.

  15. Eric says:

    @Noah – I would quibble about that being content in the same vein as what I mean, but you’ve got a fair point. However, their entire product is still just balancing and maintaining their cards. They don’t have any dungeons or so on. (Some card games DO have those; the WoW CCG has raid packs, for instance, but not M:tG AFAIK.)

  16. Eric says:

    @William Newman: I agree with everything you’ve written there. I think every MMO I’ve ever played has some degree of “found strategy”, but I get that you’re arguing for a very high level of it.

    I don’t think finding emergent behavior is bad gameplay at all; it’s a type of gameplay that especially appeals to game designers, so we talk about it ad nauseum. But most emergent-behavior games are flops. Since you can’t plan the emergent behavior, it tends to be severely flawed in some way. In MMOs it tends to result in problems creating and maintaining content; I guess the bigger risk might be that the emergent behaviors are flat-out boring.

    If your MMO can risk that kind of danger, it means you have no funding and are thus an indie MMO. I guess Bryan Green would tell you to make that game, even though it’s likely to fail. I would say make a game with a higher success rate first. The odds of magically getting a great game that way are lottery-winning small IMO. But of course, if you get it, then you win the lottery.

  17. Eric says:

    @Bryan Green – I find your comments to be astonishingly cynical. But I’ll save my argument for a future blog post!

  18. Haversack says:

    M:tG does suffer from optimal configurations. Every new block there are a few cards or combos that make up the top ranking tournament decks in type II. Now this isn’t to say that there aren’t exceptions to this.

    Alternate win conditions, besides reducing someones life to zero, help keep M:tG interesting. I can’t really think of any ways to do alternate win conditions for MMO’s not predicated on health loss off the top of my head but its something interesting to consider.

  19. Xenovore says:

    Quote: “Class-based systems are about choice, Skill-based systems are about calculation.”

    Completely disagree. It’s quite the opposite in fact. Class-based limits player choices to specific actions as pre-determined by the designer. Skill-based allows players much more freedom to determine their own actions.

    @ Eric: I have a lot more thoughts on this, but I don’t have the time right now to expound on all of them, so I’ll just hit a couple things.

    Regarding terminology… We cannot use “class” and “role” interchangeably; they are not equivalent. All RPGs have roles; not all RPGs have classes. I.e. a class may define a role, but a role doesn’t necessarily define a class. A “class” – at least as defined by D&D and most derivatives thereof (which include 90% of the MMORPGs out there) – is a set of predetermined, unchanging abilities, i.e. the game designer has already decided what a character can do, end of story. “Role”, on the other hand, is more general. To use WoW as an example: the “tank” role can be filled by different classes like warrior, paladin, death knight, etc. The “healer” role can be filled by a cleric, druid, paladin, etc. Certain classes routinely switch roles, e.g. paladins and druids, but they are still limited to the abilities defined by the designer.

    Whether class-based or skill-based, of course players can only perform a single role at a time, i.e. you cannot typically fight and heal at the same time. But, unlike class-based, where the character is locked to a specific role and set of abilities for the lifetime of the character, skill-based allows role changes by the moment,and over the lifetime of the character. E.g. a character can fight one second and heal the next. Or a swordsman can decide later in his “career” to forget about swordfighting and become a hunter, learning archery and tracking skills. Skill-based systems allow this flexibility “out of the box” and do not rely on the typical kludgery of class-based systems, where, to provide some modicum of choice to the player, designers must resort to creating multiple (21!) classes, and players are forced to create alts if they want to do anything different.

    To summarize, I’m not arguing that skill-based systems aren’t more difficult to design and balance. But just saying class-based systems are inherently better because they’re easier to design and balance…? No way can I agree with that! A well designed skill-based systems has so much more potential to provide for player choice as well things like emergent game-play. (Note I said “well designed”; most of the arguments against skill-based systems point directly to bad design, nothing else.)

  20. Pingback: /AFK: Classy Edition « Bio Break

  21. Memphis says:

    I will just say this.

    I find it funny that anyone can claim that a classless system leads to monopolies in power. Last I checked, in games like WoW, for each class there is 2-3 uber builds per class, and if you don’t use those builds, then you die to the top players, and that build changes sometimes monthly depending on patches/expansions released.

    Champions Online has classless options. The Classless in the end tend to only win their first few matches since thats when they have the element of surprise. After that, they get their rears handed to them because the classes systems just respec to own that build.

    MMOs that require more stats and less skill turn into a game of accounting. “I have +2 Dexterity more then you and 1 more advantage point put towards my ice spike. That means I win!”

    And even then, games like TF2 with few classes, and few weapons have overpowered class issues (Engineer + Turrent control and Sniper + Bow were the current OP last I had played)

    League of Legends suffers this and all it has IS premade characters with 4 powers (and 1 passive) each. It is IMPOSSIBLE to balance a pure stats game in the end, since, statistically speaking, there will ALWAYS be a dominant stat. Whereas in skill based games, the biggest problems I ever really run into is poor, exploitable game engines.

  22. Jason Brodsky says:

    I think you may have erred in your new definition of classes precisely because it misses your main point.
    Roles aren’t the key reason to have classes. If WoW only had the DPS classes, players would still appreciate the differences in flavor & mechanics. Developers would still appreciate having distinct classes for easier balance, so players could enjoy their different mechanics and flavor without laboring under 10x differences in power.

    Here’s my definition of class:
    A class is a group of builds* that players can switch between easily.** A class system uses multiple classes to provide distinct*** builds which have sufficiently little in common that the power level**** of a class can be set by adjusting features unique to that class.

    You can have many different levels of classes in a game: warrior, two-handed warrior, two-handed-mace warrior. The harder it is to switch between classes and the more value the game gets from having a variety of classes, the more important it is that classes be balanced.
    * A build is a collection of choices about your character that determine your effectiveness in gameplay
    ** What is “easy” depends on the player. In particular, some players will see a large barrier in having to reimagine their character’s flavor–some people really want to be a fire wizard and will be sad when ice is better no matter how easy it is to switch their build.
    *** Here’s where roles come in: different roles offer a great way to distinguish classes, but not the only way.
    **** The power level of the class is typically that of the most powerful build in the class, since by definition it’s easy for anyone to switch to the most powerful build within their class. However, you typically want builds within a class to be somewhat balanced as well. Using a talent tree system or something similar turns builds into mini-classes that are also easy to balance by adjusting what’s unique to each tree: the top-tier bonuses.

  23. BryanM says:

    Bad cards in Magic exist because:

    1. It makes them more money to add filler to their booster packs. (A great reason to play pauper, that.)
    2. They’re there to balance the limited formats, Draft and Sealed Deck.

    The lowered power level slows down Limited games, eliminating 3-turn cut throat win madness. Simple little cards like Grizzly Bears not being trash, also simplifies the playfield of these formats a bit.

    Guild Wars has such a strange dev team. They create skills 2 to 8 times weaker than equivalent skills in the same attribute, just as some way to limit the power level of attributes, I guess.

    (As an aside, for a game inspired somewhat by Magic, their trainwreck “Codex Arena” feature is a mind numbing example of Missing The Point.)

    @ “Last I checked, in games like WoW, for each class there is 2-3 uber builds per class”

    That’s still 20-30 builds. The point rather is if all the skills were in the same bucket, there would be only around 2-4 best builds flat out in the entire game: Single Target DPS, AoE DPS, Control, Defense, and Healing.

    I completely agree with separate buckets than an ala carte anything goes system. Games like Titan Quest do give you the option of picking the wrong buckets, but Magic is pretty good about this: no combination of two of the five colors are completely gimped.

    @ “League of Legends”

    Street Fighter Balance, with character tiers and all the fun of not picking the bad ones that entails.

    It is completely possible to balance these games… but as Eric mentioned in one of his articles, there are many subgames to take into consideration.

    Take for example, the humble Zergling. 1 Zergling will slaughter 1 Marine. However, a battle between 24 Marines and 24 Zerglings has a very very different result. Games like Warhammer Online RvR turn into Blobwars, where liquefying people at range is dominant.

  24. Dan says:

    Eric, as a fellow designer, I was wondering if you could take up a thought experiment for me. One of the main drawbacks of role-separation systems is that they often force you into a defining role very early, aka at character creation. They rely on schemas to create expectations of play style, which may or may not work. Now, most games try to turn this into a strength (new content if you re-roll, immediately different gameplay, etc), But the classic training model has few significant choices up front, many at the end (particularly multi-month significant choices). Classes fight this model – they force players to make their most significant choice right away.

    As an alternative, consider FreeRealms, which rights flips triangle head down, giving you more roles to specialize in as you play through the game. Eve, I believe, takes a middle ground, by inducing a substantial in-game cost when you switch roles, and starting with the trainee-type early game role.

    However, it seems like an assumption of class-based games that you have to pick a class up front. As a thought experiment, in a role-based system, how would you maintain a strict separation of roles while backloading as many significant character role decisions as possible? Something different from the examples above? I doubt FreeRealms qualifies as a role-based system in your examples (since the roles don’t directly interact).

  25. Memphis says:

    necroing yes. But all I have to say to bryan is. If balance for PvP is your only concern, then tehre will NEVER be balanced. When I say there is 2-3 uber builds Per class, that is not a positive. That means “You either build this way from the start, or lose” It is infact WORSE then having 2-4 uber builds because it is easier to screw up by the billions of pieces of bad advice, instead of just looking at the consistent top builds.

    And sorry, What happened in WoW at the time of the Shamans Release? The Shaman and Paladin each had one build that owned EVERYTHING except eachother (The Paladin owned the shaman) the game is a constant fight for nerf and buff trying to balance them.

    You need to face fact. If you have a game that uses the same sttas for PvP and PvE, the game will NEVER be balanced.

    Why is it people cry balance? But then want class systems, etc. ? Even if you gave EVERYONE in call of duty for example a Dinky Pistol as their ONLY gun in ANY mode… There would then be a dominating tactic.

    Stat balancing is not a sign of weak design, but of overly clever (See: Sad/Lonely) players learning to twist, abuse, stack, and outright turn a system into their puppet.

    You can NEVER truly balance a game. The only way you can try is get rid of class games ALL TOGETHER and make the game 100% skill based. StarWars Galaxies had the right idea before they screwed up with the wookie expansion. You can stack stats all you want, but if I am a better player then you, then my badly stated level 20 will still own your level 40 Min/Max character since I am much better at the fighting part of a fight.

    Making a game pure stats means there will ALWAYS be a stat that dominates, or a tsta that can be stacked to ungodly levels. But unlike WoW, at least champions lets me look awesome while failing, whereas in WoW, that game you worship, I have to look like an IDIOT and fail if I dont minmax from the start.

  26. Dave D says:

    I’d like to make the counterpoint to your two blogs. I’m in no way saying that your wrong, I agree with the points you’ve made but I disagree with the conclusions.

    From a purely players perspective, I can give you the rules that most MMO’s follow.

    Trinity based game. Crafting will never be better than drops. Rock, paper, scissor’s PvP (anticlasses). PvP substituting for high end content. High end content requiring large teams to complete. “Rebooting” a character is very expensive in some fashion. Items/Gear get’s bound to the character in some fashion. Most drops are pointless vendor trash. Lot’s and lot’s of levels to grind through. Questing is more XP rewarding than random mob killing. Your character will be part of a faction and that faction will be opposed. Progression through the game will be artificially gated by some means, usually by your level versus the mobs level.

    Since your a designer, I’m sure that you could add a mountain of other “musts” to what little I just posted, but that’s where the problem lies. There are 54294 other games that run under that exact same formula and your trying to compete versus them.

    So why not a change from the pack? Yes, it’s a lot more work. Yes, it’s far more difficult to balance and you can pretty much forget ever having a semblance of balance in PvP (if you even bother to add it). But you could make a game that had the crafting of pre NGE Star War’s Galaxy’s. You could make one that had the twitch combat of AC2. It could have the range of skills that Champions Online has. It could have the speed of travel that is enjoyed in Anarchy Online to move over the massive spaces that exist in WoW or LOTRO. It could have the ease of teaming that is found in City of Hero’s while still allowing one to solo most of the content. Your game could allow one to keep progressing his character forever, instead of forcing a roll of a new character at some distant level.

    Now all of these things might not be possible or certainly not in budget for a game, but I can assure you that I’d pay to play such a game. I can’t give you such an assurance if you just intend to make one like the other 54294 games that are out there.