Classes vs. Open Skill Systems

A few years ago, I was interviewing for a systems designer gig. The design interviewer had been at Turbine many years prior, so we vaguely knew each other from there, but she needed to know how much I’d grown as a game designer since then.

She asked:

  • “When you came to Turbine, how did you feel about classless systems like Asheron’s Call 1 has?” My answer: “I loved them and wanted to make all games that way.”
  • “Do you still feel that way?” My answer: “No, I wouldn’t inflict that upon a game that I wanted to see succeed in the long run.”

This was the correct answer, if you’re keeping track at home, and I got the job.

AC1’s system of free-form skill purchases was a lot of fun at first and there’s no denying it. But my experience showed that they had serious problems, including:

  • Lack of diversity: there is usually one or two perceived “ideal” setups and most people choose them. This makes it quite difficult to create grouping scenarios with divergent roles.
  • Inability to predict power levels: creating fun content is already quite difficult. Adding in the notion that people of the same “level” might be wildly different in their ability levels makes it even harder.
  • Difficulty adding new verbs: it is tricky to add new open-ended skills to a system without sending everybody into a tizzy.
  • Difficulty balancing: the lack of granularity makes it very hard to fix overpowered skill combinations. The already-hard problem of balancing becomes much harder.
  • Poor expectation management: it is quite hard to teach players what they will end up doing later in the game.
  • and many more.

I hesitate to say this because the general designer-blogosphere will likely disagree with me. I’m not trying to pick a fight; I’m just telling you all how exceptionally wrong you are!

Now before you get all frothy, it’s true that each limitation of open-skill systems can be worked around. Templates help with expectation management. Balancing content requires some clever mechanisms, but it can be done. You can use skill dependencies to avoid overly cheaty builds. To encourage diversity, you can heavy-handedly add “role skills” into your mix, with some gating mechanism (such as skill-point cost). And on and on.

And it seems like many beginning designers weigh the options and go “Yeah… I can make this work!” And you can. Sort of. But you’re forgetting two things:

  1. You can’t perfectly balance an MMO.
  2. There are a million times more players than there are of you.

So players instantly find the loopholes, and you spend a vast amount of time working against them. This is still true with classes, but when players are “locked in” to their class, they can’t go too far wrong. Yes, it’s a designing with a safety net. And yes, it’s necessary, in order to keep balance issues from damaging your game’s lifespan.

(And yes, balance issues can accelerate the ruin of a game. Champions Online’s mess was… well, a mess. And it’s still pretty messy last I checked.)

Open-ended skills sure sound more fun on paper, though, don’t they? For a large number of gamers it seems like the perfect system. Who could argue that you need to take the cop-out approach? Anybody who is entrusted with designing a multi-million-dollar game, that’s who.

Free-form skill choice is much harder than classes. That’s what it boils down to. You can massage free-form skill choices in a million ways, but you’re doing extra work just to get the usability levels that simple classes have built in. Thus, there’s less time to design other features. And even worse, it means there’s less time to balance your game — yet your game needs even more balancing than a class-based game does.

The even bigger cost, the deal-breaking cost, is over the lifespan of the game. The live team will have far fewer resources than the production team, and you’re saddling them with an extremely fragile beast. If your live game desperately needs new game systems, it really really sucks for the designers to instead have to focus on enormous player-hemorrhaging balance issues. And trust me, your balance issues will be enormous at first, and they will be a lot more damaging than classes ever are.

What I Am Saying Is Not Obvious. This is a case where theory and reality have a wide gulf. That’s one reason why this is so controversial. If you design systems on paper, you are quite likely to decide that the difficulty between classes and open-ended systems is pretty much a wash. That’s not reality, and that’s why this was such a good interview question.

It’s a trick question in some ways, because if you’re a reasonable sort of designer, you don’t often go around declaring game mechanics to be unworkable. You waffle. You know that if this is true, then that is true. If this is the case, then that might work. But here, the waffling would have shown a lack of experience. Sure, there are scenarios when a free-form system is the right choice. But that would be a pretty contrived scenario compared to the vast number of MMOs where it’s the wrong choice.

The Exception?

Would I ever make a game with free-form skills? Oh sure. Indie ones, ones where I knew that failure was a possibility and that was okay. I would also assume the game has these qualities:

  • largely solo-oriented,
  • a strong PvP focus (or some other “generic content”, as opposed to tons of hand-crafted quests)
  • any group content is designed for a “gang of very similar bad-asses”, not a large group with diverse abilities and roles.

Depth Illusions

Those qualities I listed happen describe EVE and Asheron’s Call, by the way. They are both successful games, given the history of MMOs. But the thing you may not realize is that both  games vastly under-delivered on their promise. At first it seemed like there would be a million crazy character builds, but it turns out, nope: just a few. It’s all smoke and mirrors, and to add insult to injury, you have to use an add-on or a web-page to remember the “right” way to level your character, because the facade of free-form-ness actually gets in the way more than it helps. And trust me, the developers spent a LOT of time working around the balance limitations of their skill system. That means updates come a lot slower than they otherwise could.

There are lots of other possibilities besides strongly typed classes and very generic open skill lists, of course. So there’s a continuum here. But it’s not a linear one. Moving just an inch towards “open skill system” brings massive amounts of extra work. And annoyingly, any compromise will create rabid angry fanboys who claim you’ve failed… before they even try it.

Asheron’s Call 2 was an innovative system of skill trees mixed with gating skills that led to particular roles. Ultimately it inherited the major flaws of both skill trees AND classes, and was quite expensive to maintain, yet it didn’t get “credit” for how open it was, because… drumroll… “it had classes”.

Guild Wars has very open-ended classes that took a ton of time to balance, and it shows in how the rest of the game was rather underdeveloped at launch. But despite a masterful class design by very experienced designers, they fell into the same trap as everyone else does: there is just one “right” way to play most classes, or maybe two; the vast number of theoretically-viable options are ignored as inferior.

Was it really wise to put that much effort into all these unused skills? Just to let people think the system was “open ended” for a short time? Maybe so — that’s a defensible argument, anyway, one which the developers of Magic: The Gathering have in fact explicitly defended. But if you accept this premise, you need to know what you’re doing here: you’re intentionally creating an illusion of complexity that will wear off in six months, tops. If it takes you twelve months to make it, you’re betting a LOT of resources on that illusion. For a card game, okay. For an MMO, you’re losing a lot of crucial polish time on a transitory feature.

The worst part about Guild Wars’ story is that their design was clearly a resource trade off, trading polish for a larger-than-normal number of viable character builds… but they still get “points off” from irrational fanboys who think it would be way better without those stupid constraints. Frustrating. (You can’t really make those people happy because they don’t really know what they’re talking about. Don’t try.)

Really! Free-Form Is Harder Than It Looks

I admit this isn’t quite an open-and-shut case. But too many designers (especially armchair designers) think you can just write this item on your MMO’s feature list. “Oh we’re gonna have a classless skill system, and crafting, and auto-leveled PvP fighting, and…”

"Vaaaaporwaaaarrrrrre! And if you do ship the game, very depressing retention!"

Whenever I see this in a high-level design, I hear a little sound in my head. It sounds like a raven cawing “vaaaaporwaaarrrre“. It is a very sad sound.

It’s VASTLY more expensive, in the long run, to create and maintain a free-form system.

Of course, in theory you can make it work and still have time left over for everything else. In theory you might also be the 1 person every 25 million years who ends up being Batman. Why not play the odds?

You know what, if you’re a master RPG designer I’m sure you can do this. It’s true that no MMO has done this particularly well — their “classless” systems tend to collapse into just a few viable templates, which are no better than classes — and it’s true that games that focus on the openness of their skill system suffer from a lack of polish in other areas. But the history of MMOs is still young.

Just do me a favor and don’t try this as your first MMO design. I’m pointing out a trap here to help you avoid it.

This entry was posted in Design. Bookmark the permalink.

45 Responses to Classes vs. Open Skill Systems

  1. Alrick says:

    I’m not a designer, I’m a player, so my comments reflect my perceptions as an end-user of designer’s work. But still…

    EVE’s open skill system’s highest attraction for me is not the variety of possible templates, it’s the easiness of re-spec. I would be completely fine with classes if it would be easy to pick up new classes beyond the first one without having to go through levelling from scratch all over again. In EVE, if I want to become a dedicated medic, I spend 3 months’ worth of training and some money for a ship, and voila! During these 3 months I am not forced to go through noob part again, I do not create alts. The downside, of course, is tat I have zero experience in playing a medic, but luckily, EVE combat is dumbed down and slow enough to make learning new roles easy.

    Actually I wonder if putting heavy restriction on changing classes or ever re-speccing within a class has more to do with genre expectations than anything else. What’s so bad about giving every player the ability to gain access to multiple classes and allowing him to switch from one to another more or less on the fly?

  2. Gathrog says:

    Although I can’t disagree with your knowledge on how hard it might be to create an open class system. As a long time gamer, who has played many MMOs, I can say that my two main reasons for leaving most of the MMOs that I wasn’t happy with were:

    First the horribly bad “Faction” system that forces you to work together with half the server and fight the other half. The best PvP and fun I’ve had playing an MMO was Lineage 2. This is a hard statement to swallow for many who have tried Lineage 2, since there were many reasons to be turned off by the game. But the open faction system was so much fun that I endured all the other atrocities forced upon me by the developers (and gold sellers) just for the social uniqueness and personal PvP that having no factions brings.

    Second of course would be the “Class” system. The multiple development issues that you introduced, like I said before, I can’t argue with. I have not tried to develop an MMO, so I can’t pretend to know what’s easy/cost effective and what isn’t. But as a player I can say that on at least three occasions (WoW, Aion, Warhammer Online) I spent many months leveling a class only to discover how much I hated it, or how much I wanted to play another class. In the case of WoW and Warhammer Online I forced myself to deal with my chosen class for years. In the case of Aion I flat out quit a month after reaching max level.

    If there was an easy way to switch your class without having to spend many more months starting from scratch, it would have been a massive improvement. I’m not talking about extra starter gear or a weak “Apprentice” type system for 10% faster leveling. After playing a game for months nobody wants to go back to learning how to use strike again for 12 hours. The system also doesn’t have to be as easy as paying a fee and then magically you are another class either. But a system where you can change your class while staying in higher tier content is more then reasonable in my eyes.

    Something like this would also help developers be able to easily target what classes need balancing. As you said in your post, the vast majority of players WILL play the most effective builds. Even going so far as to research what might be best and follow it to the letter just to find out. If players were given an easier way to change classes you can see what classes are most effective and fun, and what classes are not, just by looking at the player population of each class. No system, as you also said, will lead to a perfectly balanced MMO. But the current “Class” standard has PLENTY of room for new ideas and development.

  3. Pingback: Classless is a pain in the assless « Stylish Corpse

  4. Joshua H. says:

    This general topic is one I’ve been on for a few years now but the inclusion of EVE in your post has made me re-evaluate my stance. Not that I think skill based systems are a good idea but, more, why they’re a bad idea. Comparing EVE to other skill based systems (defined as a game where a character develops individual skills instead of fixed packages) you see a large divide.

    Where most games allow you to pick your skills but only X amount, EVE lets you pick, in theory, all skills. In practice, given time, a single player can gain skill unlocked “proficiency” in pve, pvp, gathering, production, research, exploration, and trade. It isn’t maxing out skills but a single player can engage in a wide variety of activities without the need to respec or reroll. The larger part of where EVE differs is in that it only allows you to do most things one at a time. Trade, research, and production happen in the background, more or less. The active portions of the game are essentially class based where your class is defined by the equipment you choose. Equipment choices are always situational but the same applies to class based games. Even class based systems with heavily locked down gear sets have situation-limited skills. Picking the right spells to cast at Mob X, for example.

    As I now see it, the problem isn’t between players choosing a locked skill package vs a free-form skill package but, instead, that players in the free-form systems are often given the impression that they can create their own “class” and rock on with the best of them. For a skill based system to not spin off course, the designer needs to inject faux-classes over which they have control. Let players see a role they can fulfill and give them the freedom to spend banked skill points their sword and shield and holy skills and unlock the Paladin class that they can then take into battle.

    The balance question isn’t about skills and progression, it’s about what freedom the players are or aren’t given when combining those skills. Do you let a player make their own classes or not.

    The only way to really save players from using the wrong skill (be it a non-viable trained combination or shooting a fire elemental with a fireball) is to remove all skills. And then you get Minecraft.

  5. Steve says:

    FFXI did a variant on the class-based system that I haven’t seen anywhere else yet. Mabye it’s too hard to do, and that’s why no one else has tried it.
    Except (kind of) EVE, in that EVE lets you skill into multiple specialties, and then you can switch between them by piloting and arming different ships.

    Your default account only gave you one character slot (you could get more by paying an increased monthly sub fee per slot). But, you could unlock all of the classes on your one character, given enough time and inclination.

    You still had to level up each class, so the EVE comparison holds here.

    FFXI, on top of this, had a sub-class system, so that you could have a primary class (e.g. White Mage) and then a secondary class (Black Mage), both active at the same time, and so your primary class would receive buffs from your secondary class, and you could use some of the skills from your secondary class. (I’m probably remembering the details wrong, but I played FFXI a long time ago, and I’m doing this from memory.)

    Anyway, to min-max this system, you’d want to have a secondary class that complemented your primary, like the WHM/BLM combo example. Having Warrior as a secondary to WHM (or vice versa) wouldn’t provide any benefit, since the two classes are so different.

  6. Conor says:

    My introduction to online RPGs was a MUD that allowed you to pick one “primary” class and two “secondary” classes out of seven different options. You’d have strong skills in your primary class, and much weaker skills in your secondary classes (so that they were basically for utility, not power). The classes were all very focused — the Scientist had only healing & repair skills, the Assassin had only melee damage and stealth skills, and so on.

    These days I play WoW, and watching the designers solve balance problems by giving similar skills to everyone (no more unique buffs, all melee get a short-cooldown interrupt, all DPS get a weakish self-heal, etc.), I can’t help but think back to that multi-classing system where the answer to “my class needs the ability to heal itself!” was “then go secondary Scientist, dummy.”

    Do you feel that multi-classing solves any of the problems of completely open skill systems? (Or single-class systems, for that matter?)

  7. wufiavelli says:

    Big fan of the class system with the skill facade like eve. Keeps everyone happy oddly. People are easy to trick.

  8. I’m going to be very curious how Trion handles these issues with Rift. I suppose that’s technically a class-based system, but it might be just wide enough to hit them with the worst of both worlds.

  9. Eric says:

    Great comments here!

    Gathrog: I agree that classes work better when you have a way to switch to other classes. I don’t think “but… you’ll get FLAVOR OF THE MONTH!” is a very good counter-argument about why people should be locked in… it’s one thing for designers to draw a line in the sand and say “here they are: the eight classes of our game!” and it’s quite another to say “everybody should play each class in equal number, dammit! We’ll keep balancing until that’s true!” The former is good for expectation management and content creation; the latter is just being anal-retentive. But it’s common.

    Practically speaking, it’s tough to provide a high-level class-respec feature in a pure class system. In a skill system you can let them slowly shift their classes, so that they are “gimped” (AC term) for a month or so but they can still limp through high-level content. In a class system, you’re either one class or you aren’t, so it’s harder to find a way to give class-switching any weight: it seems like it has to be instantaneous, or else you have to devise some elaborate system. “Well in the first week you lose feature X but you gain feature Y. Then the next week…” Lots of fiddlin’.

    I would say the right answer is just to have a full-on class-switch option that’s instantaneous, but has a long cool-down. But I could also get behind cleverer class-switch mechanics, haven’t really brainstormed the possibilities there.

  10. Pingback: Psychochild's Blog » Stay classy

  11. Eric says:

    Conor – my take on “major” and “minor” class selection is that it’s still pretty much a cosmetic thing. By that, I mean it doesn’t really buy the designer much freedom and it creates a lot of templates that don’t work, so a lot of that work is going toward promoting the illusion of depth. It does typically increase the amount of template options, but generally not by much more than having more classes overall.

    Let me explain why it doesn’t buy you much: you start out by saying “Okay, anybody who uses Tank as their minor class gets ability X, Y, and Z”. But then you find that Priests who have Tank as their minor are overpowered, pretty bad. Do you: A) nerf priests overall (bad!), B) nerf Tanks overall (bad!), C) tweak what powers people get when they take Tank as their minor?

    It has to be C. That means making fiddly, very specialized decisions: either certain major classes get a different set of minor skills than others, or, if it’s uniform, then the minor skills are different in implementation from the “actual” skills — the Minor version of “Taunt” might be 30% weaker than the “real” Taunt.

    In the end you’re making a lot of fiddly specialized skills, and knowing that most of them aren’t going to be viable. Nuker+Tank results in “meh” for most games. So you’ve still got the problems of poor expectation management (newbies are going to assume any combos work, cuz why else would you code them?) and increased balance headaches.

    It’s not a BAD solution; I just think of it as being a little further down on the “classes vs. skills” curve, and ends up with some of the problems of both. If done well, like in Guild Wars, it can give your game a fair amount of divergent setups. But I don’t think it’s a “cheap” solution in terms of the amount of headache you get for the benefit.

  12. Troy says:

    I would have to question your assertion that PvP content could favor open skill systems. I would think that the intense competition that usually results would put substantial pressure on builds far from the ideal, although you would still have people searching for new ideas and options.

    Could you talk more about how you think PvP would suit open skills?

  13. Tiber says:

    I hear what you’re saying and I understand the balancing nightmare. But as a player who’s drawn to RPGs largely for character building, and who will play unique builds even if they’re sub-optimal as long as they’re interesting, what exactly do you have to offer? Sure, you can reduce the number of options and make a pretty balanced game, but one of the main draws of an MMO is supposed to be freedom, or at least the illusion of it. In your opinion, where do you draw the line between maintaining balance and allowing players to make choices, even ill-advised ones?

  14. Eric says:

    Troy – the logic there is:

    -In the short term, PvPers will try many different combinations in order to find the best combo. This group tends to be especially efficient at finding broken combinations.

    – Since I mentioned it was a primarily solo-oriented PvP game, that’s okay: they will gravitate toward one uber class anyway, say a tankblasterhealer.

    – There will still be myriad bad skill combinations and lots of players who ultimately feel they need to reroll. In a PvP game, though, this is inevitable if you have any power dichotomies at all (and you will). So the lack of classes isn’t doing TOO much more damage to the game than the players themselves will do to it by feeling the need to be optimal.

    – In my experience PvP players seem more willing to reroll anyway.

    – Since it’s a PvP game, you don’t have to worry about creating content that is balanced for all the possible types of players.

    Really the last point is the most important one. You could replace “PvP” with “farming monsters” as the main mechanic and it would also work out more-or-less. What DOESN’T work is when you have WoW-style quests, because it is ridiculously hard to create quests that are a perfect challenge for all the different power levels your game will have.

    This is especially true early on because YOU (the designer) won’t even know what the most broken skill combination is — so the content you create before launch is inevitably going to be a cakewalk for the optimal skill build. That means that in about 12 months when even brand new players are savvy enough to read the internet before playing, the content you created will be balanced for NOBODY and you’ll have to revise all of it, which is time consuming and low-reward.

    (It’s “low-reward” to the live team because your players never say “yay, you revamped the lowbie zones!” they say “Damn it we need more high level zones WTF are you doing?!”, because they don’t care about your newbie retention, they only care about content for themselves. They don’t realize that your retention may be so bad that it’s threatening the lifespan of your game… players don’t give you credit for thinking about anything other than themselves directly. So you want to avoid revamping old content for as long as possible, because it’s a time sink during which you aren’t giving your core playerbase much to chew on.)

  15. Eric says:

    Tiber – there’s plenty of room in classes for variety without exploding the balance issues. Sometimes this is done with talent trees, sometimes just with vast quantities of classes. These approaches tend to result in more playable skill combinations than freeform skill sets anyway.

    But you’re right that it doesn’t FEEL as wide-open. However, it’s largely an illusion. What you are seeing in a skill-only system is a facade: most of the characters you create won’t be fun to play and you will not be happy with them for long. (Well, some people are happy to keep plodding along at a fraction of the speed of others, but most players become frustrated with this. Their retention tends to be shorter than a player who is moving through the game at the speed intended by the designers: they are not having as much fun. This is true even for people who will keep their gimpy character rather than reroll them: at some point they feel too attached to the character to restart, but they are missing some of the game’s magic that keeps people hooked.)

    For naive players there’s no complete way around this problem: the illusion will keep winning. Games that shout “hey, we have a free-form skill system!” will keep drawing them in, even if they end up disappointed later. The best I can offer this crowd is a game with lots of cosmetic differences — think City of Heroes, where everything from a ninja to a robot mercenary can use the same skill set, and mostly they just look different. For some players that is enough to do the trick. For other players, that isn’t a good enough illusion for them.

    For savvy players that really want mechanic differences in order to feel distinct, I think the answer lies in creating games with broad class experiences. Asheron’s Call 2 had 21 classes (!) ranging from classic wizards and rangers to people who telepathically controlled swords to fight for them, to bee keepers who flung nests at their foes, to stone-throwing giants with a penchant for poetry. Jeckyll & Hide alchemists. Elementalists in control of the weather. Blood-drinking witchdoctor-healers. On and on. And each of those classes had a dozen or so choices to make. (Not major choices, but somewhat significant ones.) When the number of possible classes is overwhelming like that, you get much the same vibe as you do from an open-skill system: you feel somewhat unique.

    This approach has down-sides — making dozens of classes isn’t exactly cheap. But it’s a helluva lot better than flat open skills in terms of maintenance and balancing. It may seem bizarre but creating two dozen classes IS cheaper than creating an open skill system.

    As an aside, this is where D&D 4th edition went. There are now dozens of classes (across a dozen expensive rulebooks), each with many distinct variations to pick from. From the outside looking in, it probably seems like 4th edition has really diminished the potential number of different characters you can play with this ham-fistedness. But in practice, it’s actually created a LOT more playstyles that are fun and viable. (D&D 4ed has a lot of problems and I’m not trying to downplay them, because they are significant. But lack of character diversity isn’t one of them.)

    The trick to having lots of classes (without going insane) is to keep track of the important verbs and not cross them. If “being a tank” entails five specific abilities, and “being a healer” needs these other abilities, and damage-dealers need these others, then you never cross the streams: you never give full-power versions of damage-dealer abilities to healers, or healing to tanks, or so on. (Low-power versions of those abilities are fine, because they are usually comparatively useless and just for flavor.)

    So you don’t cross the main class-defining abilities. Instead, you allocate a large collection of verbs as being “up for grabs”. Then you tack on a bunch of these extra verbs to each class. Tank + crowd control, Tank + buffing, Tank + debuffs. Give each one a wild flavor: “this guy gets his power by cutting himself, he’s the emo tank. This version uses magic butterfly cocoons he got in hell. This one’s power is completely illusional and sometimes he stops believing, and his powers stop working for a second.” Whatever. Go crazy and be unique. The results tend to be pretty awesome, and you’ll get enough differences that most people will find something they find fun.

    I think the broad range of flavorful experiences AC2 offered was one of its best points, and something I would love to see other class-based games use.

  16. Pingback: Defending the Lack of Class…

  17. Tiber says:

    Talent trees work for me. I was mostly commenting on the fact that you didn’t care for multi-classing systems either (and having played a game with those problems I know where you’re coming from). Note that I won’t necessarily play a gimped class, just one that’s not considered the best class but has some redeeming feature or can do something other characters can’t.

  18. Pingback: The Skills of EVE « Aim for the Head

  19. William Newman says:

    Eric wrote “But you’re right that it doesn’t FEEL as wide-open. However, it’s largely an illusion.”

    It’s not always that it’s an illusion. I’d say in many cases the opposite is more nearly true: the problem is not that the choices are illusory, it’s that players being able to make real choices torpedoes the game design. Various design elements in mainstream successful games conspire to produce perverse outcomes when players are given meaningful choices. When ordinary players have to make meaningful decisions it follows (given the other design elements) that most of the outcomes are un-fun.

    Partly it’s a very broad set of design elements related to making game challenges dramatically satisfying. E.g., anything RPG-like tends to go to any lengths necessary to limit real-world strategies involving concentration of force: “so that was bad, so then we regrouped, and we came back with five times as many of our buddies and a week’s production of satchel charges, and we stomped the place” is not the kind of dramatic resolution that designers want to encourage, or even permit. The contortions required to prevent this from happening tend to cause a lot of collateral damage to logic in the rules.

    Another problem is that if players don’t have the information to make decisions, they’re generally going to make bad decisions, and this badness is an un-fun kind of frustration. Even with the best will in the world, documentation is tediously hard, and creating fun systems with interesting properties which are simple enough to document is deeply hard. So within, say, WoW’s system of class and talent and profession choices, you can easily get poor outcomes by not knowing which timeouts interact with others, and Blizzard doesn’t even try to give you enough documentation to figure that out before you commit to your bad choices. Similarly, you can get bad outcomes by making choices that are inefficient given drops you can’t possibly have found yet, and you can get horribly bad outcomes by interactions of your choices with stupid boss tricks (utterly arbitrary insanely powerful boss immunities and special powers and behaviors) or interactions with abilities you could only learn about from deep in other classes’ spellbooks (e.g., polymorphing into an animal is significantly less powerful when it leaves you uniquely vulnerable to hunters’ ability to cause fear in beasts). Something similar was visible in Diablo II: e.g., play guides have to warn mages not to do the obvious strategy of getting really good at one school of magic because long after you’ve made irreversible choices based on that strategy, you will start discovering that various important monsters are absolutely immune to any one school of magic (but, it vitally and undocumentedly turns out, never immune to more than one school).

    Incidentally, the two problems I’ve picked on interact: note how common it is for games to fiddle around behind the scenes so that things stay dramatically close even when you’ve made such good or bad decisions leading up to them that now things should be seriously unbalanced for you or against you. Giving players usable information about how the world actually works would interfere with this fiddling.

    I think an important part of the practical importance of class systems is as a workaround for a complex of problems including the ones above. By constraining the players to roll through the game on rails on one of the dozen or two dozen platforms the designers have specifically planned and tested for, the designers make it practical to succeed in working behind the scenes to keep gameplay from being totally destroyed by how their design keeps players from making good choices about meaningful decisions. Limit the choices which are important to ordinary players to choosing between no more than a few dozen meaningful alternatives, fiddle until each alternative supports a fun outcome, et voila, you can have a fun game even when your design practices produce so much internal inconsistency and so little valid player-visible information that (ordinary) players can’t (short of major out-of-game efforts like Elitist Jerks or the walkthroughs for older games like Diablo or Elder Scrolls) engage in sensible decisionmaking.

    Incidentally, I’m not sneering at the other design elements here. Existing games are fun, and it’s largely an unsolved design challenge how to create comparably entertaining games without using design elements which interfere with sensible decisionmaking. My point is just that these things are tied together. Letting players make meaningful choices is a deeply powerful probably-good thing. Compare, e.g., to putting personal computers on the Internet, similarly a d. p. p.-g. thing. When you do a deeply powerful thing, it can end up connected to other design decisions you have been accustomed to making (about, e.g., multitasking, security, and reliability) in ways that can make it tricky to decide in the abstract whether “putting a computer on the Internet is a good thing” and which in practice mean you may want to do a lot more engineering than just adding a network card and a TCP/IP driver stack to CP/M.

  20. Troy says:

    I see, that does make sense now. It’s not that PvP favors open skill systems, but rather that tuning set-piece encounters for challenge in a skill system is much more difficult.

    And hopefully people will auto-sort themselves for challenge if given the tools to do so, as that tends to happen with competition.

  21. Noah says:

    I’ve always enjoyed GW because they did have the balls to do the two class system; it’s terribly difficult to balance and sure people still go with the cookie cutter main/sub class but, for me, (like someone said earlier) I like to try out the different combinations even if they’re considered “sub-par”. It leads to the depth of the game, even if you only have 20 levels. I would gladly take 20 levels across 3 characters versus 60 levels with one character. The differences are the former gets 3 completely different verbs than the latter.

    Another example that is interesting to look at is Atlantica Online; they do a terrific job of creating a “classless” system. The game never gets the credit it deserves for the design albeit the game itself isn’t my cup of tea. Yes, they still have the cookie cutter builds, but you’re still left with the options of doing other things that still work.

    Lastly, for the players (like myself) who enjoy trying and testing the different options, even if you tell us it’s all an illusion anyway, we’re still going to try it–and like it. It’s almost like telling a kid not to stick the penny in the socket, sure you can tell him that it’s going to hurt like hell but for the rest of his childhood he’s going to want to stick that damn penny in there and see what happens.

  22. Pingback: Classes, Trinity and Balance, Oh, My « Tish Tosh Tesh

  23. Chris says:

    As a player, I’ve come to expect that any skill/class leveling system is largely dependant on the most current provided content, and that forcing -change- on the playerbase over time is a necessary evil due to min/maxing and mudflation, and this is only on the PvE side of the equation. Add PvP into the mix and adding constraints becomes the second necessary evil, especially if “balance” is to be maintained among the min/max crowd. So I would have to ask; what is the actual goal of designing a skill/class leveling sytem in a game where the illusion of freeform gameplay is to be fostered? Does transparency rule the roost, or do contrivances(constraints) attempt to get buried or hidden in the details?

    The thing that I, as a player, hate to discover, is getting to the end of a journey and realizing that I’m in a game of one-upmanship with the developer and not other players.

  24. Pingback: Best of the Week — 17-23 Jan ’11

  25. Eric says:

    @Troy: well sure, you could say it that way… if you wanted to be succinct and to the point about it. :)

  26. Eric says:

    @William Newman: I think you’re confusing verisimilitude in with other concepts. It’s certainly true that the “real world” could be better represented in games, but that’s an entire separate ball of wax. The biggest problem with this philosophy is that games aren’t realistic. In your example, if you lost against the dragon, you would be dead. You wouldn’t come back, ever.

    And you wouldn’t have any magic, either. What sort of “realistic” rules of thumb would a player assume about magic? I guess it would depend on what they’ve read. Vance novels, Lord of the Rings, Harry Potter? You can’t make everybody happy, or even most people. Your game instead needs to have an internal logic.

    The other counterargument is that games often DO have a lot of verisimilitude but don’t get credit for it. In EQ2, if a boss monster stomps your party, you can get fifty friends online and go murder him to smithereens. So? Why would you?

    Just as if you got into a fight at a bar and called in thirty buddies from college, you will win but the rewards aren’t exactly going to be very high. The games permit it just fine. But just as in real life, there’s effort vs. reward to be considered.

    Similarly, real life is pretty bad at helping you make good decisions. Consider how many people wish they had gone to school longer, or wish they hadn’t stayed in school as long. Like the Diablo 2 wizard who thinks powering up one skill is the obvious best plan, real life did not necessarily go according to plan.

    Actually, real life SUCKS at managing expectations. Games do it better already, by a long shot. There’s still lots of room for improvement, but real life doesn’t provide any special examples on how to do that.

  27. Pingback: /AFK: Forumophobia Edition « Bio Break

  28. Wayne says:

    Having played AC for way too many years, I realize the foolishness of the skill system there. Sure people talk about all the different skills and how you can combine them to make so many different character types, but it boils down to a variation f a handful of templates for the majority of players. The exception being the stubborn ones and the role-players ( I fall into both camps). I thought AC2 was going to do it right, but it was just really messed up at launched. My opinion they were just getting AC2 right when it shutdown, oh well.

    I like the idea of a handful of classes and talent trees within those classes that offer very specific skill sets. Take 3 basic classes, range fighters, melee, and magic, then allow people to specialize. As always specialize in one area gives you a weakness in another area. Not sure how this would work or how popular it would be long term.

  29. Fig says:


    I’ve been reading your blog for a little over a year now, and your posts always leave me feeling satisfied that at least someone in the gaming industry is still in touch with reality. On this point however, I have to respectfully disagree with you, in spite of your industry experience. I do agree that classless systems are generally more difficult to implement, but I don’t believe that this is necessarily an inherent design flaw of skill-based systems. Rather, I think it’s a far deeper problem arising from the way in which we design games, and the player’s role within it.

    Personally, I think this issue is a holdover from the very first RPG games, and the rise of the holy trinity of archetypes; namely the tank, damage dealer and healer. It’s no surprise to me that if you create a class-less system with skills focusing on these traditional roles, players will find the optimum builds for fulfilling them. You gave them the verbs to do so, and implicit in the game design itself (especially if the game includes PVE) is that there is a need for all of these roles. Thus, players will fill those roles in the most efficient way possible. The real advantage here is that players are able to experiment and change their character within a short amount of time, though the disadvantages can be anything up to crippling balance problems, griefing, exploits, etc. Certainly, class-based systems are much easier to design and anticipate for within this traditional paradigm.

    However, I do believe that more freeform, non-traditional games are possible by discarding current design principles. As an example, you mentioned in a previous article the problems of balancing for 2 players with identical attacks. The first one to attack always wins. Then, if you give one of them a heal, that player always wins. The simplest solution to problems like these is to break the rules. Give a player a healing spell that can only be used on other players. That way, the player still has group utility, but isn’t unbalanced in 1v1. Hell, you could combine healing with both damage and tanking, within boundaries. E.g. a nuker that does a little bit of group healing (other players only) based on the amount of damage they do. Or tanks that need to rotate in and out of tanking to heal one another, and generate their healing ability by actually taking damage.

    Personally, I think the healer archetype could quite easily be removed from MMOs entirely, and more unique group play introduced. I’d actually like to see an MMO come out that isn’t another massively multi single player game. Skill-based games are ideal for this, because they tend to mitigate the onerous task of leveling a character, and allow the player to focus on other things such as crafting, PVP, roleplay, exploration and gathering, and metagaming. The point that always seems to get lost nowadays is that MMOs don’t always have to be purely about PVP, and it’s okay sometimes for fights to be a little unbalanced. That’s one of the things that makes winning worthwhile.

    Then there’s the psychological aspect of not knowing exactly what type of player you’re interacting with. Many games nowadays go for instant recognition; that a player should immediately recognise a given class and how they play, just from the way they look and move. But whatever happened to keeping your cards close to your chest and using your last ace to pull out a win? What of learning that one sneaky ability that your opponent doesn’t expect, and using it in a clutch moment to make the difference? Some builds will always be better than others at performing given roles, but in PVP there tends to be far more variation in tactics. Even in WoW (back in vanilla) you had Arc/Fire mages, scorch builds, elemental builds, frost builds, arc/frost, etc. There tends to be much more leeway in PVP because there’s no numerical value that can be attached to survival skills and snares. You can optimise for damage or survivability, but a player with a slightly less-optimum build than you can still outplay you, and a survival ability such as a teleport or freeze is only as good as your ability to utilize it.

    This of course leads me to the other major problem with current MMO design, which is the focus on the questing and dungeons. Quests really only serve the function of helping players to level. Sure, they can convey a story, but let’s be honest here; most players don’t read quests, and really just want the XP to help them level. Similarly, dungeons are a device to create the illusion of progression, by drip-feeding progressively stronger items to the player base. In essence, dungeons will always limit builds far more than any skill-based or class-based system, because there will always be a most efficient way of doing them. Monsters can’t react to a given strategy, and so the strategic meta-game that forms around PVP never takes shape. You just go the necessary spec, and play your role until you get the loot.

    If you decreased the focus of PVE from dungeons and quests, and made it more along the lines of gathering resources, crafting, creating and destroying player owned buildings, and meta-games such as making your own tavern for people to gamble in, or even things like racing mini-games, I think you’d find that skill-based systems are actually better suited to the task, and provide for more satisfying gaming. I mean, why shouldn’t ‘end game’ start from the moment you make your first character? That’s really the only bit that people want to play anyway, so why force them to play a particular way in order to get there, and then force them to keep playing that way for the duration of the game?

    My thoughts on the matter.


  30. Pingback: More on Classed vs. Unclassed Games | Elder Game

  31. Pingback: You Gotta Have Class… Or do you?

  32. William Newman says:

    Eric, you responded to my post by writing about “verisimilitude,” games resembling the real world. Perhaps you were taking off from my “tend to cause a lot of collateral damage to logic in the rules,” and making a reasonable guess that by “logic in the rules” I was referring to a realistic simulation. However, I wasn’t trying to say specifically that verisimilitude is what’s important. My point is about something more general, that what’s important is whether players can successful reason about game outcomes. A sufficiently realistic game is one specific way to give players enough information to reason about outcomes, but it’s not the only way. (And with today’s computer tech, it’s often not a very practical way.) Simple complete arbitrary sets of rules (Bridge or Chess) and completely unrealistic regularities (if it’s in Minecraft, it’s a block) are other specific ways to achieve the “logic in the rules” that I was trying to talk about. The arbitrary regularity that D2 monsters are never immune to more than one school of magic, or that Oblivion monsters grow more powerful as your experience level increases, are other unrealistic things that let players make reasonable choices once they figure them out.

    Even when everything is known up front, e.g. Chess or Go, players can make all sorts of lousy decisions. E.g., beginners in Chess usually have to be told by more experienced players how important it is to develop all their pieces early, and beginners in Go usually have to be told how important it is to arrange for their stones to stay connected into a few large groups. So the problem of lousy decisions doesn’t go away when information for making reasonable decisions is available. The information just makes the problem qualititatively more manageable (and IMO also qualitatively more fun).

    You write “real life is pretty bad at helping you make good decisions” and “real life SUCKS at managing expectations.” I guess; but to some extent anything with meaningful choices has that problem. e.g., Chess and Go can be very frustrating. But at least they do give you a chance to make good decisions.

    Besides the practical obstacles (e.g., documenting stuff just takes time) to giving players enough information to make good decisions, it seems to me that game designers are often actively opposed to it. A while ago I was looking at Elitist Jerks trying to understand tradeoffs for those elemental shaman decisions which depend on the swing speed of weapons. From memory, my understanding of the situation they were analyzing is that before some level of item quality, it’s pretty common for weapons to be slow enough to make a decision pay, but above that quality level slow weapons don’t exist. That kind of arbitrary breakage, giving players an undocumented regularity that dozens of hours of into play at level 85 suddenly turns into a completely different undocumented regularity, seems like it might show active hostility on the parts of the designers to the “internal logic” I was trying to get at. Another example might be the D2 miniboss before Baal who is immune to cold and is discoverable by the player only when the player has trapped himself with the miniboss in a small confined space that the player can’t get out of by normal means. Someone apparently thought it was dramatic/fun/exciting/whatever that when the cold-oriented mage first has a chance to discover that kiting can’t be relied on, the revelation should be as lethally surprising a blindsiding as possible, at a level of experience where getting more experience to belatedly invest more points in remediation (buying strength for armor, endurance for hp, and non-cold magic skills) would be extremely time-consuming given that the miniboss stands between the player and all activities that would efficiently give higher-level xp.

    In some cases, designers may be opposed to giving players enough information because it’d be embarrassing, revealing a part of their game design that’s stupid and un-fun. (Oblivion levelling of monsters comes to mind.) Other things like the arbitrariness of shaman loot drop patterns by level seem to be because the designers’ priority is not meaningful player decisions, but preserving lots of arbitrary degrees of freedom for designers to adjust arbitrarily so that players’ decisions largely cancel out in order to be less meaningful than they look.:-| And in some cases like the D2 miniboss, I don’t know how to explain it except by designers thinking that concealing from players the information they’d need to make globally reasonable decisions, and going out of the way to leave players trapped in a particularly deep hole for having made various locally reasonable decisions given the information they had had at the time, is a design goal in and of itself.

    I think in order to make a meaningful answer to questions related to “do meaningful choices tend to contribute to fun?” you probably need to break it into at least two questions, “do meaningful choices tend to contribute to fun given game designs which hide the information needed to make reasonable decisions” and similarly “…given game designs which reveal the information needed to make reasonable decisions?”

  33. Xenovore says:

    @ Fig: You nailed it. Could not agree more.

    @ Eric: Definitely a good read. But it seems derived more from personal preference than anything else…

    Also, this boggles me a bit: “Asheron’s Call 2 had 21 classes…” You say this like it’s a good thing, but exactly how is 21 classes easier/better to implement than 21 skills? What if a game only had 5 skills, no classes? Wouldn’t that be easier to implement and balance?

    My point is: I don’t think the issue is really about “classes vs skills” but rather “good design vs bad design”. With skills, the design challenge is to provide meaningful and useful actions for each skill, i.e. does the game design make each skill necessary? Will players want to use every skill? And most importantly, is each skill fun?

    The solution looks straight-up to me — if a skill is not particularly useful or fun, throw it out! Just as you would throw out a class that isn’t useful or fun. For example, if your game only has 3 potion types, or making potions is boring and tedious; then it’s pointless to implement a Potion-Making skill. On the other hand, if there are 57 potion types, and they are fun to make, then it makes perfect sense to implement the skill.

    Of course you’ll still have some skills that will be more useful (and/or popular) than others, but really the same can be said for all the typical features of MMOs: classes, skills, talent trees, gear, quests, etc. Any of those can be poorly implemented, not just skills.

    Also, there are two key problems with class/level-based MMO systems: 1) Content gets out-leveled and becomes extraneous, and 2) low level characters cannot play with high level characters. I think a well-designed skill-based system can go a long way to mitigating these problems.

    It also definitely depends on the over-all game type. A simple class-based system work well enough for a themepark-type MMO; but for an open, sandbox-type MMO, a skill-based system is more suitable and has far more potential.

  34. Eric says:

    Xenovore – well… for starters, you can only pick one class at a time. Presumably you can pick more than 1 skill. I went into more specifics about why classes are better than skills in the post after this one.

    I don’t believe that skill systems have shown any particular value in helping people group together or in keeping content from becoming unvalued, except inasmuch as your skill system is actually just classes (as defined in the next post).

    To be specific: If your game lets you buy lots of skills but only use one roles’ worth of skills at a time, it is a game with classes. It is not an open skill system. EVE Online, for instance, is a classed game.

    (I didn’t realize it when I wrote this, but it’s obvious in retrospect from one of the pingbacks here; I talked about it a bit more in the post after this one.)

    I think a lot of your argument is just a problem of conflicting terminology — my problem with using bad terminology. Please check the next post on the blog, I think it will clarify things.

  35. Eric says:

    @Fig: I agree completely in theory. In practice, you’re hand-waving over the hardest part. What you’re hoping for is a free-form skill-buy system that lets you choose from multiple group-combat roles, or pick and choose elements of multiple combat roles. (If a player can only have one combat role at a time, then it’s a classed system, regardless of how people get their abilities.)

    So, about picking parts of multiple roles. First, those roles need to be defined by the designer or they simply won’t exist. Combat roles very rarely appear from user innovation; more typically, combat roles DISAPPEAR because players figure out ways to play without them.

    Thus AC1 intended to have large grouped encounters with a healer, buffer, tank, ranged dps, and up-close blaster DPS. These roles disappeared as players found ways to do without them. Obviously AC1 wasn’t very well designed in this regard. But you’re saying it’s easy.

    It seems like you’re suggesting that it’s pretty damned easy to design a system so that there’s multiple group-combat roles, and let people pick and choose parts of them, and it will all work out. That’s not easy. That’s REALLY HARD. The more dissimilar (and thus typically the more fun and compelling) your combat roles are, the harder it is to let players randomly pick chunks of them.

    The point of this post (and the next one) is that this is harder — mathematically provably harder, even. It’s not just my experience talking. If you look at it logically there’s no way to see this being simple.

    If you design a game for six player groups, with six fun combat roles, brilliant ones — the best ideas ever — and then let people pick chunks of them so that they can effectively have two roles covered at once, what are the odds that the players are really going to need a group of six? What are the odds that they’re going to need three or four, instead? Very high. Since you probably balanced the game for five- or six-man groups, these smaller groups are going to level faster and get better rewards than you intended. Like, almost twice as fast and twice as good… or if you do your job poorly, 6 times as fast.

    They’ll fly through your content and then bang the drum for more. In the mean time, people who didn’t pick mutated class layouts will be plodding along. Now it’s time to make new content. Do you ignore those slow 6-man groups and just make new content for the 3-man groups? Or do you keep making content that 3-man groups find useless? You’re stuck, you can’t make good content, you can’t make the game compelling for the various play styles.

    Fixable? Yes. You can ingeniously craft six combat roles so that any person can take any subset of them and end up being roughly equivalent in importance to other players. Easy? Hell no! It’s going to take you a year or more to get it ballparked (in other words, so that the efficacy of leets vs gimps is only divergent by 200% or 300%). Your game will never be balanced, but that’s not important. What’s important is if you have them ballparked so your content is useful. If your game goes live before that point is reached, you’re quite possibly ruined.

    The reason people keep using the old “trinity” roles is because we know them well enough that we can riff on them easily. We have some understanding of them that makes it at least somewhat practical to make healer/tanks. If you make up entirely new roles, you won’t even have ANY practical experience to do this. I say mixes of the “trinity” roles tend to double up — so a six-man group could become a 3-man group. If you make up six brand new roles and let people mix them, all bets are off. In many games a single build will end up being able to do six-man content.

    I’m not saying we shouldn’t innovate roles! PLEASE make your next game with all new group combat roles! (That is not the hard part.) But don’t let people pick multiple chunks of them at once. (That is the hard part.) Save it for the sequel when you understand these roles like the back of your hand.

    As mentioned in the next post, whether players buy their skills from a tree or a list or have them handed out by the Class Fairy is not really relevant to my argument. If your game has well-defined roles, and a player can only perform one of those roles AT A TIME, then your game has CLASSES by my definition. Gussy it up any way you want, ‘s cool.

    Finally, I have to poke at your argument that dungeons and questing are just there to progress your play. *ALL* elements of the game are there to progress your play. You’re noticing the raw game mechanics more easily than you’d like; players who see the mechanics nakedly will call it “a treadmill.” Other players call it “the game.” There is no game that doesn’t have content of some sort. What am I missing from your argument? What is it you want to see? Are you saying you want dungeons and quests to be harder to find? Less accessible? Rarer, so most of the gameplay is undirected monster-stabbing?

  36. Eric says:

    @William Newman: Okay, sure. I can see that some players will find it lots of fun to perfectly stat their characters, and that they are frustrated by the lack of documentation about game systems that keeps them from doing that.

    Okay. But … how is this related to classes vs. classless games?

  37. William Newman says:

    Eric, you and I seem to basically agree about the main thesis of your article that classless RPGs are very hard to balance, quite possibly so hard they are absolutely impractical to balance for multiplayer play. But I thought I saw you adding in effect “and even if you could balance your classless RPG, it would still not be fun, because players making real choices make real mistakes and that’s not fun.” (That was how I interpreted the sentence just after the passage I quoted in my earlier reply: “What you are seeing in a skill-only system is a facade: most of the characters you create won’t be fun to play and you will not be happy with them for long.” It sounds to me like you think that is some combination of inevitable and unfun.) That’s what I was trying to disagree with. I think if somehow designers could get the balance issues under control (perhaps in part by dodging them with approaches you mentioned like preferring generic content to hand-crafted content), a flexible system would be a lot of fun.

    Many people find fun in playing with the choices in skill-based single-player RPGoids like D2 and Elder Scrolls, despite severe balance problems. (Admittedly Elder Scrolls is not a great example because it’s not that much more flexible than a system with a modest number of classes, but I don’t know that all that many computer RPGs so I don’ t have many candidates.) And going outside RPGs, games with lots of choices and emergent properties (e.g. Minecraft and Dwarf Fortress) have an impressive amount of entertainment value even when wildly unbalanced. To the extent that there’s a pattern of players being unhappy because “most of the characters you create won’t be fun to play” I think it’s not some absolute fundamental problem of games that let players make bad choices, but a problem caused by secondary things like your problem of designers not being able to anticipate balance issues which follow from the outcomes of the choices (which I agree is very hard, perhaps completely intractable), and my problem of designers and implementors not letting players find out what they’d need to know in order to make reasonable decisions (which I think is widespread and relatively tractable).

  38. Brian J says:

    Ummm… has anyone heard of or played Uncharted Waters Online? Makes me think this article was written with a certain “Western” perspective on MMO’s.

    That game has skills & classes, and they have no issues with balance. Every class has skills and no one person has enough points to learn all the skills. Basic skills can also be added regardless of class and players can choose to change classes if desired through in-game mechanics.

    It’s open-ended, it’s skill based and it’s class based. But… it was also designed by a Japanese company and released in the Asian market well before it ever made it’s way to Europe or America. Seems they were ahead of the “skill vs. class” debate that other developers got bogged down in.

    Other than that, yes. I agree 100% with the arguments against skill-based systems. I loved Ultima Online, but everyone always leveled the same skills for the same end-game results. Classes at least give the option of multiple end-game results depending on your class.

    It’s impossible to pick out one portion of an MMO without also relating it to the rest though. For instance show me one class-based MMO where everyone hasn’t created an “alt”. Most of the skill-based games I know only allow one character. Alts simply mean people have 2 characters experiencing two aspects of the game. But now, this also impacts the economy in the game. Who needs to buy from a crafter if their alt can make it?

    Also show me one class based game that had the population spread out in the entire world. I always think it’s such a waste when I play a game years after it comes out to find the “newbie” zones a vacant ghost town. All the work that goes in to developing them is wasted once a character levels past the zone.

    For the year I played WoW, once I finished the quests and the mobs no longer gave anything significant, I would rarely go back to that zone. All the labor that went in to making the zones were wasted once I no longer needed anything in them.

    Skill-based games tend to avoid this – if the skills are done right that is, where level caps prevent everyone from learning everything and you need to juggle the points if you want to change roles. In order to practice a skill, you need easy monsters to beat up, regardless of your other stats. If you “forget” a skill from lack of use, you need to head back to these easier areas to work it back up. The population of the game world is always spread out, not just concentrated in the end-game zones. This also means new players joining will usually see a populated world, not a deserted wasteland for 90% of the game.

    Skill-based games aren’t 100% perfect, but neither are class-based games.

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

  40. ThatRickGuy says:

    I agree with the author so long as we are using the concept of “Ability = Effect”.

    I press the 5 button, my character performs ‘Ability X’ which does ‘500-600 damage’.

    In an classless/unstructured skill based game, that would be incredibly difficult to balance. Heck, even with classes and strongly typed skills, it has proven to be an increadibly difficult to balance job. Just look at WoW, they’ve been at it for years, and after each expansion comes months and months of rebalancing.

    If you break from that concept though. Instead of basing output on the ability, base the output on the character’s progression. For a grossly simplified example, if a character is 50% progressed through the game, their damaging ability would do 500-600 points of damage. Whether that is from a bow or a spell, a slow two hander or a fast shive, insta effect or a dot. Effectively, all damage is capped at a normalized point. All healing, all threat, all damage reduction, etc… gets capped based on the progression of the toon. A level 20 rogue and a level 20 warrior and a level 20 mage when all played perfectly, would all be performing right at those caps.

    Then again, I’m in favor of a fully free formed skill system. Where you’re not picking from a pre-defined list or ‘leveling’ an ability, but actually building custom abilities. For instance, if you learn how to swing a sword, you could ‘progress’ it by investing skill points in being able to swing faster, or to use less energy, or to increase your crit chance, or base damage, or add a dot, or add a utility effect (stun/disarm/knockdown/etc…)

    If you go that far from the structured class/skill/ability farm, it has to be done formulaicly, as the author says, trying to balance such a system of ‘abilities’ would be increadibly difficult. But if you don’t balance the ‘abilities’ and instead balance the characters, the issue appears to become much more simple.

    I’m no MMO developer. I am a software developer though. And I have some friends in the indie game field. Maybe I’ll put together a demo engine of how I would envision such a system to work and see what they think of it.


  41. Pingback: | Superlevel

  42. William B. says:


    I have a question for you:

    Why do you think developers are so obsessed with the idea of balance?

    1. Either never change a ‘class’ at all.
    2. Allow people to switch / rebuild.

    In my opinion, one of the big reasons that players leave games is having their hard work nullified. People play games from minutes to years, yet the second you create your character, there is a certain attachment that goes along with it.

    When a games constantly have tweaks and adjustments for the sake of balance, you rarely end up with the same character that you started with.

    I feel that developers get an obsession over balance and ultimately end up shooting themselves in the foot.

    Most players, play MMO’s for freedom, exploration, achievement, etc. but when you’re not attached to the progression that you’ve made, you lose a lot of those opportunites.

    Let’s take AC2.

    AC2 had it’s (un)fair share of class / balance changes throughout it’s lifetime (understatement), but the best thing about AC2 was the options that it provided.
    Most of the players didn’t mind balance issues, with respecs there was always a flavor of the month… as well as the counter to that. Things were allowed to flow and change.

    A huge detriment to player retention is stagnation. When a player can no longer tweak their character, they grow bored or angry (gear doesn’t count). A system like wow where you just auto gain skills and then have a few tweaks with talents, is so soulless and uninspiring.

    I would rather be ‘underpowered’ then be the same as everyone else, even if being different is just an illusion.

    I guess what I’m saying is please don’t forget the spirit of games. Yes, you’re in the business, but your customers aren’t, they are out there looking for adventure, not some perfectly even, balanced, standardized, road map.

  43. Pingback: This Week In Video Game Criticism: Hidden Morality And Self-Imposed Permadeath | Video Games

  44. browolf says:

    I think the difficulty to balance comes more from complexity than structuredness. But Unstructured games are inherently more complex. Wow I find to be structured yet complex because of the vast array of choice within specs. They’re always having to tweak the balance and they seem to suffer from the law of unexpected consequences a lot. Whereas ffxi which I play most is structured but a lot simpler. They can add stuff without it affecting the balance. Even when they increased the cap which gave access to more powerful subjob abilities it didn’t break the game. They even lowered the level of some abilities to get them into top level chars. Every time wow adds levels they have to rebalance all the specs.

    the worst part of wow was having to pay gold to respec. Taking away the freedom to figure out what works best because I couldn’t afford the rising gold cost. thus forcing me to look online to find out what are considered the most optimal spec. Futher forcing me into a specific play style. I really hated the way wow gives options on one hand and limits the ability to use them how I wish with the other.

    It seems to me that ffxi’s subjob system (a version of which used in ffxiv and copied by runes of magic) is a particularly clever way to create options without adding complexity and you don’t get to choose your skills. On paper there’s a lot of the same characters, the difference in reality is by what equipment you use and how you use the limited skillset every other person like you has. for instance Ninja class before level 39 works best with an “off-tank” however having one that knows what they’re doing is a hit and miss affair. Or black mages have a set of spells called elemental debuffs. they’re really helpful but plenty of blms don’t really know what they’re for and never use them.

  45. VIPer says:

    In general i disagree with this topic.

    BTW i have some questions :
    -Why players with no role don’t create role for themselves?
    -Where is exactly hardness in balancing? Is this for group PvE only or for all multiplayer form of gameplay?
    -Why class system have more fun when skill-bying?

    As i understand-this problem most for predefined quests. And for some players quest may be too hard, for some-too easy.
    But in PvP, for example, i really can’t see big problem in this.

    Sorry for bad English.