Psychochild's Blog

A developer's musings on game development and writing.

20 January, 2011

Stay classy
Filed under: — Psychochild @ 5:24 PM
(This post has been viewed 7604 times.)

After making a new year's resolution to post more often, Eric over at Elder Game has decided to re-animate the corpse of an ancient discussion about Classes vs. Open Skill Systems.

Oh, look, this topic again. But, let me try to look at this in a new light.

Eric goes through the old litany of reasons on why class-based systems rule and skill-based systems drool: design is hard, players over-optimize, skill-based systems once had a drunken party one night and got arrested a decade ago for urinating in public. Er, something like that. After reading so many of these, they tend to blur together. Oh, and the insinuation that only immature designers see classless systems as appealing is added for good measure.

The ever lovely and talented Ysharros over at Stylish Corpse posted an interesting reaction where she laments Eric's proclamations and takes a look at what it might mean for her unhealthy obsession with an upcoming game. (Personal to Ysh: Get help. Get professional help. Get better professional help. Remember to take your anti-psychotic drugs.) Anyway, a recommended read to get a more player-focused view.

Harmful terminology

One of the big frustrations in discussing game design is imprecise terminology. What is a "class"? In MtG, as Tesh refers to in a comment on Stylish Corpse, is "Blue" a "class"? Well, if we're going to shoehorn that game into the MMO paradigm, then it would be.

But, is that distinction useful for discussion? Not really, because the term "class" has a lot of emotional baggage with it. By advocating classes Eric is obfuscating his message and potentially harming those immature designers he is trying to warn since they'll see "classes" and make a whole host of default assumptions.

False dichotomy

The flaw is seeing the option as two separate, binary values: class-based or classless (skill-based). These terms both have their own baggage and assumptions. Class-based means that you pick the same archetypes as every other game and implement them in a strict way. Skill-based means you have a laundry list of skills that players can pick as they choose and thus create the ultimate tank-mage jack-of-all-traes. In this false view, classes are obviously superior.

But, instead of being binary, the options are more of a continuum. (Well, really, it's a vast space of possibilities, but let's not go too crazy in a single blog post.) Let us pick some new terms so get away from the emotional baggage: let's use the term "structured" to describe a character's development. Strict classes in EQ1 are the most structured example, with UO's free-form skill system (and the original tank-mage game) being the least structured example in major MMOs. Note that AC1 is a very unstructured game when looking at Eric's list of complaints about how hard it was to balance the game.

But, as usual, the best answer lies between the two extremes. As I've mentioned before, Meridian 59 achieved a level of balance by providing some structure. Each "spell school" had a certain theme (like MtG's colors) and spells were ordered in ascending levels where you had to learn enough of level 1 spells to progress to level 2, etc., up to level 6. You could only learn so many levels of spells/skills based on your stats, so that limited how you could build your character. By imposing some structure it was easier to balance the game and it addressed the issues brought up; but, ultimately, it allowed the players more freedom than a strictly structured class-based system.

Funny enough, I think a good modern example of less structure comes from DDO, based on the granddaddy of class-based systems. The flexible multi-classing rules in 3rd edition D&D allow for some interesting character options. The enhancement options in DDO allow you to create some interesting characters. My favorite character is a level 6 Cleric/8 Sorcerer. I'm able to be the primary healer in an on-level group as well as throwing in some nukes arcane-only buffs like Blur and Haste. Admittedly the character has less maximum power potential compared to a super-focused single-class character, but the flexibility allows me to have fun with my group. But, it's widely acknowledged that multi-class options in DDO are difficult for players to tweak properly, leading to a lot of "gimp" characters if not handled well.

Overwrought danger

Let's take a look at the issue that skill-based systems are "hard to balance." The truth is that all games are hard to balance, and even relying on the crutch of class-based design is no magic bullet. Okay, who here thinks that WoW Paladins have always been completely balanced compared to every other class? Anyone who thinks so, take a look at the advice I gave to Ysharros above. But, the question is: did it make the game any less popular or fun? Sure, there might be some people who got upset and left over Paladin issues, but most would agree WoW is still a highly successful game.

What about the issue that players will over-optimize? Again, they do the same thing in class-based systems. Let's look at WoW. Want to play an all-around powerful character? Play a Paladin (or Death Knight in WotLK). Want to solo without a care in the world? Hunter is what you want. So, is optimization acceptable only if it was provided by the developer instead of discovered by the player? But, let's not pretend that people still don't feel the need to follow a template; the details of how to "build" the character in a structured class-based system like WoW has its own specialized site.

The argument that "fewer people will use underpowered skills" doesn't hold water compared to classes, either. The most popular WoW classes are played twice as much as the least popular ones, at least in raiding. Warlocks, in particular, are also unpopular in PvP, so is it a waste to put effort into them. On the other hand, some people still play them.

These problems aren't unique to less structured games. I think the real benefit is that its easier for inexperienced players to not know or not care about them in more structured systems. If a new player rolls up a Warlock in WoW, they may not care that the option is unpopular, or that they need to spec a certain way to be invited on raids, or that it's not the most powerful option. They can have fun with it and perhaps enjoy the trappings of the class more than the calculations underneath. And that, I think, is the real design advantage of highly structure classes, in that players can recognize classes form other games and enjoy the flavor as they learn about the game.

Achieved goals

Ultimately, as a designer, you have to ask what your goals are for the game and use the right tools. Perhaps a highly structured class-based system is what you need. Or, maybe you want a less structured system. The first step is to identify what your goals are and then pick the tools that will help you achieve those goals.

--


« Previous Post:





32 Comments »

  1. Two posts on here in two days?

    I'd like to apologize to anyone who thought they had until 2012 for the apocalypse.

    Comment by Psychochild — 20 January, 2011 @ 5:24 PM

  2. Pfft. You just want my TSW beta spot. Don't think I can't see straight through you. (Now back to reading the actual post.)

    Comment by Ysharros — 20 January, 2011 @ 5:45 PM

  3. Weird timing, I was just replying to comments on this post when your pingback came in. I think you're wrong. :)

    - Here's how I draw the line, and I've made this argument before so I didn't think to repeat it: assuming that your game's "roles" perform a certain set of key verbs, then you have a strict class system if those key verbs are restricted to certain roles.

    - By this same logic, a system becomes more free-form the more major role verbs a given player can obtain. A game is completely free-form if they can obtain a large number of major role verbs. My contention is that such a game is orders of magnitude more difficult to manage than a class-based game.

    - Now you can (and should) sprinkle on lesser verbs (which might include poor-functioning variants of major verbs, for flavor) but these don't change the "classness" of your system.

    - The "classes versus open skills" continuum is a large one, but you're making it sound linear. That's not true. At one end of the curve is regimented EQ1 classes with no divergence: incredibly easy to maintain. At the other end is AC1 with no structure: incredibly hard. We're in agreement here, I think. But where it breaks down is when you suggest that any point along that line is a linear choice between the two. Not so. When you leave the class area your complexity ramps up VERY quickly. To use the boring old "trinity" roles, when DPS can heal (at meaningful levels of effectiveness), you're going to have a hell of a time balancing.

    - Giving players flexibility WITHIN a role is different, I'll grant that, and there's a ton of ways to gussy that up so it feels like meaningful choices, but isn't. I think it's pretty easy to give Healers several fun play-styles while still keeping them within a balanced range. This idea was outside the scope of my post though.

    - DDO is extremely difficult to play without help from websites. My first three characters were entirely gimped because the game provided me with false expectations of what I can do. They do have a fun game system but 1) there's not much actual variety and 2) it's hell on newbs. I did in fact quit the game because I was gimp and not willing to reroll. I don't think their design was a good choice in terms of player retention, even if there are a few moderately-viable middle ground characters as a result.

    - The old "but class X isn't played much compared to class Y" argument really doesn't hold much water with me. The assumption is that this is because one is more powerful than another, but that is usually not the case. Developers follow the crowd, and the crowd goes where the fun is (at least in a game as massive as WoW... with a smaller audience you might get more grognards fast). I would say the Warlock is less used because the playstyle just isn't much fun for as many people. That just points out that WoW has some sub-optimal classes. Nobody would argue that every class or skill in a game needs to be equally important. However, WoW makes sure all those classes are within a pretty tight continuum. And that is VERY important for expectation management.

    - You're really glossing over the power differentials by saying that WoW Paladins have been overpowered in the past. That's true... they've sometimes been as much as 300% more powerful than others, at least as far as PvP goes. However, in AC1, the best build was 50,000% more effective than the worst build, and maybe 5000% better than the "expected" builds. Creating content for this sort of range is impossible. You have to stop what you're doing and try to address the issue, or else accept that some gimped people are simply going to quit playing those characters, which is not exactly good customer service. It is in fact the worst possible customer service.

    - When I DM, I work very hard to make sure everyone is having fun. If somebody is really gimpy, I eventually have to take them aside and say, "Look, I can tell you're frustrated. Why don't we give your character X, Y, or Z?" That sort of personal service is not possible in an MMO. The frustrated players LEAVE. That's why I'm always hesitant to break out tabletop game examples. Human beings can help fix up serious problems at the gaming table. Computers only do what you tell them to, even if what you tell them to do is give certain players a bad time and make them quit.

    - At the end of the day, we both agree that players will find a small set of roles that are viable. No matter what your system is, they will pick the paths of least resistance. Given this, and given the tiny amount of time a typical MMO has for developing skills, it just makes more sense to shear off lots of those extraneous paths from the start, rather than implementing a lot of complexity that falls by the wayside within six months of the game's launch.

    - UNLESS, of course, you're trying to pick up the group of players who are routinely fooled by the promise of an open skill system, and who often don't even complain when it turns out 7/8ths of the skill complexity goes unused. (From what I understand of M59, most players would consider M59 to be class-based, btw. Players don't perceive a vast variety of options. It's classes or not classes.) But if you're catering to those players, you're choosing to spend a lot of time on this facade of complexity. If you're confident that it's worth it, so be it. I'm not. I think they can ding you on the lack of class variety all they want but if you had an extra six months of time to make content instead, that will keep your people around longer than lots of additional class complexity would.

    - I used indie games as my example of when a purely unclassed system might be reasonable, but actually, I think most indie games should focus on getting a clean set of roles working first, then add in various flavor options as they have time. There's just not enough time for perfection.

    Comment by Eric — 20 January, 2011 @ 6:18 PM

  4. Hmm.

    Read the rules to the hero system. Goes through my mind in this discussion. And I'm too tired to expand on that at this point in time :)

    Comment by unwesen — 20 January, 2011 @ 6:42 PM

  5. (And just to clarify, the biggest difference in the imbalance between Paladins and the god-king AC1 PvP character was that Pallies could be fixed within a few weeks. Fixing AC1's imbalances was a deep, extremely difficult problem because of the lack of structure. These two types of imbalances are just millions of miles apart.)

    Comment by Eric — 20 January, 2011 @ 6:43 PM

  6. The "hard to balance" question is less about the structure of the choices and more about the number of meaningful choices. There's a perception, which may or may not hold for any specific game, that a "classless" or otherwise flexible system is going to have more total choices, because this type of system gets marketed as offering more customization. The reality is that a class-based system can get into just as much trouble (e.g. Blizzard's call to slash the number of talent points and place strict limits on their distribution).

    That aside, I maintain that DDO is an illustration of the dangers in more flexible systems. Could you get to approximately the same place in terms of versatility with a Rogue 1/Cleric X with UMD for arcane wands and scrolls? It seems like just about every build that isn't staying pure for maximum casting DC's tends to include a level of Rogue splash, in large part because you can count on an infinite supply of wands and scrolls in a way that you wouldn't in pen and paper. See also the overwhelming expectation that characters will be built with at least one Toughness feat, heavy fortification, and preferably evasion. This forced Turbine to balance content around these capabilities, turning anyone who failed to get that memo into one-shot bait once you get to higher level content. In a class system, it's easier to say that the tank gets all these things and everyone else needs to stay out of the fire or whatever.

    Comment by Green Armadillo — 20 January, 2011 @ 6:51 PM

  7. Ah, one of my favorite topics... of course it would pop-up on a hell week at work. Oh well, I'll sneak a few quick comments in, and hope to get back before it's hopelessly stale...

    I certainly agree that the "field of options" goes well beyond a mere dichotomy, or even linear, progression. The "meaningful choices" concept is really more toward the core of the issue. The easiest game of all to "balance", after all, is one with no options at all. It's more a question of "how far out from the center (zero-point) do you have the resources to venture?"

    Also, I like the perspective of "key verbs" defining roles in play. It sounds like a solid method of analyzing a design for the tank-mage problem, IMO.

    All that said, I don't personally believe that start-of-game is necessarily the best place to force a character to choose their primary (if not only) role for their potentially multi-year experience in your MMO game, another very typical assumption of what generally gets referred to as "class-based", and a big part of what us in the eternally present "skill-based" dreamer crowd is against (my perspective). If my memory serves, both Raph Koster and Dr. Bartle have at instances over the past half-decade suggested ways (equipment-based is the one I recall... warrior has a sword, mage has a wand, can only wield one at a time, and it takes time to switch) that perhaps allowing people to switch roles with a high level of exclusivity serves many of the same balancing purposes... at the tactical, "encounter" level at least. (And in a typically combat-centric game, does anything else matter much?)

    Eh, enough molotovs for the moment, I guess. Hopefully be back later....

    Comment by DamianoV — 21 January, 2011 @ 7:47 AM

  8. Your last point, about new players, is probably the one that is the most enduring. By that I mean that balance problems can be worked out, eventually, or in some way mitigated. But the experience of new players is critical. A class system helps them to have some foundation to stand on. It's a reference point to help them from getting lost in a vast world of math and stats, most of which will be useless. While this is only recent, the addition of "these are your best stats" tip on the character creation screen means that new players aren't wondering if maybe agility isn't such a bad idea for a warlock, since dodging and armor are good for anyone, right?

    Comment by Klepsacovic — 21 January, 2011 @ 8:07 AM

  9. P.S. On the concept of allowing switching of roles mid-game: Superhero MMOs are exploring this territory already, obviously... CO and DCUO both have concepts along this line.

    P.P.S. I always wonder how much resonance there is in these discussions with the old saw about "programmer experience"...

    -=-

    A business owner was complaining to his programming-savvy lunch buddy that he couldn't figure out how to gauge how experienced a software engineer was until it was too late.

    "It's actually pretty simple, " his buddy replied. "Ask him to look at a complex project, then ask whether it's possible to add some feature."

    "What? How will that help? I don't know programming."

    "The general tenor of the reply is all you need. If the programmer essentially says "Sure!", without much reservation, you've got a relative novice (aside: or an incurable optimist). If they say "Not a chance!", again without much reservation, you've got a moderately experienced prospect, and if they say "Yes, but it will cost more and take longer...", you've got a deeply experienced individual."

    "I don't understand."

    "The novice hasn't been bitten much (yet) by blowing projects due to trying to shoehorn in additional work under unmodified timeframes. The intermediate hasn't lost a lot of opportunities (yet) due to being so inflexible. The expert gives you the information you need to decide whether it's worth it or not, and covers his own derriere while doing so."

    Comment by DamianoV — 21 January, 2011 @ 8:39 AM

  10. Defending the Lack of Class…

    [...] Stay Classy at Psychochild’s Blog [...]

    Pingback by The Rampant Coyote — 21 January, 2011 @ 9:58 AM

  11. Ah, well, now my brain's all awake and rested, so I might as well write down some more thoughts.

    From Eric:

    Here's how I draw the line, and I've made this argument before so I didn't think to repeat it: assuming that your game's "roles" perform a certain set of key verbs, then you have a strict class system if those key verbs are restricted to certain roles.

    I really like how you phrased that, couldn't have said it better myself.

    To use the boring old "trinity" roles, when DPS can heal (at meaningful levels of effectiveness), you're going to have a hell of a time balancing.

    This is the sort of conclusion I disagree with, though. Not that I'm saying it'd be simple, but I'm saying it'd be entirely achievable.

    The reason I feel that way is... kind of easier to explain in a roundabout fashion. I've previously worked in a job designing and implementing peer-to-peer networking protocols, which included designing mechanisms by which nodes can cooperatively decide whether or not another node they know is eligible to perform some critical role in the network, such as e.g. reliably answering questions about where other nodes might be found. Those mechanisms are pretty much the bit in P2P networking that are still badly understood, much researched, and very different from implementation to implementation.

    The secret sauce, though, is to have a function that maps various statistics you gather for nodes into single value that ranks nodes against each other. Transfer that concept to ranking characters against each other, and the statistics you're dealing with are character stats, skill bonuses, etc.

    Finding such a function is fairly tricky; the key point is to understand that the statistics you feed into the function mean very little on their own. What matters are compound values from several stats, e.g. stat1 * factor1 + stat2 * factor2, etc. Just about any RPG system that I've seen so far deals with those compound values at some point or another; in many you'll have something like an offensive bonus derived from your characters strength, agility and weapon skill or whatnot.

    You'll end up with a number of such values that have actual meaning to the game, and it doesn't matter quite as much whether a player increases one such value by using better gear, or by having better innate abilities such as stats or skill bonuses.

    The thing is, these values exist whether or not you have a class-based or a skill-based system. They exist whether you're aware of them, or not. Find them, or better yet start desiging your system by defining them first, and use those as the key verbs rather than "class" or "skill", and you'll be looking at the right place to balance any system. The overall rank of a character will be derived from combining those verbs in a weighted fashion.

    Your balancing act is then the answer to the question of what factors to apply to individual stats in the above pseudo-formula. But you'll already know the importance of each skill or stat, because you're aware of all the factors in the system. Now playtest and tweak those factors. Or run a few AI battles where a genetic algorithm mutates the factors, but to be honest I just love that idea, I wouldn't know how useful that actually is. Might help in finding some rough balance, with playtesting, public betas, etc. being required to fine-tune it.

    That sort of approach worked pretty well when dealing with computers with unpredictable configurations, by starting out from a few sample configurations and extrapolating running a number of tests, and tweaking things over a number of releases. I'm positive it'll work for a game system, though it probably will be a little harder to do.

    Comment by unwesen — 21 January, 2011 @ 10:08 AM

  12. Classless is a pain in the assless

    [...] Psychochild – Stay Classy [...]

    Pingback by Stylish Corpse — 21 January, 2011 @ 11:06 AM

  13. The Skills of EVE

    [...] after reading follow ups by Ysharros and Psychochild and the comments on all three posts, I still think that EVE’s sort of gear controlled skill [...]

    Pingback by Aim for the Head — 21 January, 2011 @ 12:02 PM

  14. Eric wrote:
    I think you're wrong. :)

    Wrong about what? That I have balanced what I consider a skill-based system? That a good designer uses tools to support their game design goals? I didn't think I said anything too controversial here, it was mostly a refinement of what you've said before, but with my personal experience in dealing with balance issues in a PvP game.

    I thought you were presenting a false dichotomy in your original post. You think I'm ignoring everything besides a linear scale, although I did write, "(Well, really, it's a vast space of possibilities, but let's not go too crazy in a single blog post.)" I think a good discussion needs to limit the space described, but not too much. I felt your "classes > classless" post was too simplistic; on the other hand, trying to discuss the whole space of character development would make my already lengthy posts stupidly long. So, looking at a continuum between the points you defined is useful.

    Part of the problem is the terminology, which makes it seem like a binary setup and is why I introduced the new term "structured" to describe the various options along the continuum. I think this is more useful, as it shows that the end-points are probably the least useful.

    I also think limiting things too much does a disservice to the less experienced designers who read your blog entry. I'd rather have them think about alternatives rather than simply telling them that they must design classes otherwise the game will be impossible to balance. But, I think it is also useful to warn them that a totally free-form system has a lot of perils as well.

    DDO is extremely difficult to play without help from websites.

    Actually, it's rather easy. My better half and my good friend both play it without referring to outside websites. I've referred to websites, but that's mostly because I'm a game designer and I wanted to see what the game had to offer beyond just rolling multiple characters. I've not followed a build suggestion on any site yet.

    Both my GF and my friend are experienced 3rd edition D&D players and this is the first MMO for neither one, so that might be part of it. Now, I will grant you that they're not going to develop characters that are within the top 10% of power in the game. But, that's not been a problem so far; we've been running 2-3rd level Elite quests with four characters and one hireling pretty easily. (Yeah, it gets harder later, I know.) But, I think that's part of the problem here: we're assuming the dominant player has to be an Achiever (from Bartle's classifications) and needs to maximize power. For me, the DDO system tickles the Explorer part of me as I've put together builds that are fun to play but not likely to get me spots in the top raids. I didn't mind re-rolling my characters, and I have about 5 characters I play on a "regular" basis depending on my mood and what others are playing.

    Again, I think the issue here is that limiting the scope limits the useful answers we can get. Is DDO a good first game for people with no paper D&D experience? Absolutely not. But, it fills a useful niche for people who can appreciate the game. Just because it didn't tickle your fancy or because you didn't want to play "gimped" characters or because it doesn't have WoW levels of subscribers means the game is without merit.

    That's true... they've sometimes been as much as 300% more powerful than others, at least as far as PvP goes. However, in AC1, the best build was 50,000% more effective than the worst build, and maybe 5000% better than the "expected" builds.

    Oh, if we just want to talk about PvP, you should have played a Feral Druid against a Beastmaster Hunter. The hunter is infinitely more powerful with a fear, a pet, and ranged damage. A feral Druid had no chance. Not sure if it's still that way after the Cataclysm talent revamp, but I expect so since Hunters still have all those abilities.

    But, consider this about AC1: is that imbalance because of the skill-based system, or poor design of the skills? As far as I know, AC1 was developed by some dedicated but inexperienced game developers. Would a more experienced hand be able to reduce the effect of that power disparity? I say yes, but I suspect you feel differently. The problem is that once a system is set, it's hard to make the big changes needed to fix poor design decisions that were set in stone. So, it's practically impossible to "fix" AC1 now.

    I would say the Warlock is less used because the playstyle just isn't much fun for as many people.

    I might buy that if Warlocks had always been on the bottom of the heap in terms of popularity; however, Warlocks used to be a VERY popular class back at launch. In raids, it was viewed as a waste to take a Mage along instead of a Warlock since the Warlock's damage was so much greater and more mana-efficient; the only think a mage brought was the ability to sheep humanoids (which weren't very common in raids back then) and free/water food (which you could get them to make beforehand). I remember when Warlocks were the absolute masters of PvP because they could fear and put several DOTs on you. Even if you managed to kill them, you'd often die from the DOTs afterwards. Plus they stacked Stamina since they could turn hps to mana easily, so they had a huge hit point pool for a caster. They were killing machines.

    The first round of massive changes I remember in WoW were made to the Warlock class to nerf it HARD, and I suspect that initial experience turned a lot of people off from the class. The designers might also have a hesitation to beefing it back up because of the negative impressions so long ago. Also, Death Knights were supposed to be "part Warlocks" (with a few of their abilties) and given how much they've gotten beaten by the nerfbat lately, it's probably not an easy task for the developers to balance that type of gameplay it out.

    You might be right in that it's not as much fun, but I'd argue that's because for Achievers power = fun, so a less powerful class is going to be less fun.

    (From what I understand of M59, most players would consider M59 to be class-based, btw. Players don't perceive a vast variety of options. It's classes or not classes.)

    If choosing between the terms "class-based" or "skill-based", I'd say Meridian 59 is skill-based. Your character's power comes from the skills you possess, not from a set collection of powers. But, it's not totally free-form, which is why I like the term "structured" to describe the continuum. I think it's the structure that was applied to it that makes it possible to balance. One cannot simply cherry-pick the hot skills du jour and have a big character. The system also requires players to make choices which limit other skills they can choose.

    And my experience with the players is different. They see the system as being flexible with many possible builds. Yes, there will be holy wars over what build is best, and why some set of skills are "gimp", as you have with any game that allows choice and where performance is meaningful. But in general there was a nice, wide variety of ways to build your character in M59. The grave sin that M59 commits is that you have to set some aspects of your character at creation and you cannot alter them afterward. That's a design issue that should be avoided.

    (And just to clarify, the biggest difference in the imbalance between Paladins and the god-king AC1 PvP character was that Pallies could be fixed within a few weeks. Fixing AC1's imbalances was a deep, extremely difficult problem because of the lack of structure. These two types of imbalances are just millions of miles apart.)

    Funny that WoW's Paladins weren't "fixed" in all the years. But, actually, I think the real issue is that Paladins were intended to be "easy mode", as per a WoW developer's talk slide I can't find currently. Perhaps the real design challenge isn't to design in balance, but to try to use imbalances effectively to achieve your goals.

    Comment by Psychochild — 21 January, 2011 @ 5:01 PM

  15. Yes, MMOs are doomed. People play them like a leaderboard instead of trying to have fun (besides a lot, if not a majority, of voiceless people that don't care to post about it).

    When I see classes VS skills I don't see a game. I don't see fun. I wish I'd be able to get people to play MMOs in another way but somehow it seems we only hear about passionate rage touching balance and unfairness.

    When I first played SWG I was a ranger with some pistol. Gimped. I liked it and had memorable (and impossible) moments. Others didn't because they were comparing themselves to others (and PvP didn't helped). These others were my friends and guildmates so I ended up alone not having fun. So yes MMOs are doomed with this type of leaderboard thinking.

    Oh wait... I think I posted in the wrong post...

    Comment by Dave Toulouse — 21 January, 2011 @ 8:32 PM

  16. (roll eyes) Yes, I think you're wrong that designers should use tools to solve problems. Designers should all code in assembly. Better yet, punch cards.

    No, I think you're wrong about the "false dichotomy" you discussed. Leaving aside terminology (which I agree is always difficult in these things) I don't see a continuum of choices: either you gate people from having multiple role-defining abilities or you don't. Binary choice. If you don't cross the key verb sets, then everything else is flavor gravy. If you do, then you have hell maintaining the validity of your roles and the power levels of your content. It's really that simple.

    But it's true that a PvP game that focuses on small-numbers encounters (typically 1v1) like AC1 (and Meridian 59?) is the best-case scenario for allowing multiple roles at once. That's because you don't really need that many roles. You're "balanced" when you accept that the best one-on-one murdering build(s) are the power point. Tada, balanced.

    The original AC1 designers naively assumed they'd be able to make content for five or six people with very distinct roles that would work together. In practice, the best they could do was make content for one or two bad-asses that could do lots of things simultaneously, or five or six people who could gang up to zerg rush a big monster. Okay, move the goal posts and call it balanced. Tada! But that doesn't work for a game where you CAN'T move the goal posts that far. Which includes most games that utilize canned content.

    Regarding AC1, it's certainly true that they exacerbated the problem with poorly-designed skills and costs. But as long as you could be a full-power healer plus a full-power blaster, there was never a chance for a broad range of character roles. (And frankly the biggest problem was the lack of live team resources which kept them from addressing ANY game imbalances for over 18 months. At that point, you really CAN'T change things.)

    To apply my "classes" terminology practically, AC2 is a great example of a "classes" game, despite having no official classes. You certainly didn't pick any class when you started the game. Instead, your race choice locked you into a dozen skill trees. You could pick from many of these at whim, which provided you with a great many utilities and secondary attacks, but you could only get into one of the six specialty trees available to your race.

    My experience is that nobody ever called AC2 a free-form game; instead they simply referred to the number of specialty trees as the number of classes. (Even the Asheron's Call 2 wikipedia page, and the few remaining AC2 fansites, break content down by "class".) This is in spite of the fact that there were occasionally viable character builds that didn't use ANY of the specialty trees.

    Players seem to make up their minds right quick about whether a game has classes or not; some implementations can fool them and make them confused about specifics (does Guild Wars have 10 classes, or do the major/minor classes mean there's 100?) but a "classless" system is generally easy to spot, regardless of how you present it to players.

    Regarding WoW: it has its own big kettle of fish. WoW's balance issues should have been pretty solvable, but they were exacerbated by a different problem: the desire to make each class have unique gameplay mechanics. (I'm referring to Combo Points vs. Rage versus Mana etc.) This keeps similar classes from being in the same continuum and makes comparing power levels extra difficult. But that's a separate blog post I should do (writes down notes). The rest of their problem comes from a penchant for rewriting classes instead of repairing them, tossing all previous balancing efforts out the door again and again. (But I've already written that blog post many times.)

    I forgot that DDO has new templates that help keep you from being overly gimped. That's new since I tried it. But DDO already pissed in my cornflakes when I tried to enjoy the game, so I have a really hard time imagining going back. You don't get take-backs with the first impressions your game makes, and that's kind of my other point.

    Comment by Eric — 22 January, 2011 @ 1:06 AM

  17. @Dave Toulouse: I think a lot of us have had experiences enjoying a game even though we were horribly gimped... until at some point we stopped enjoying it. But most people stop playing any given MMO after some amount of time; this is the "initial retention" period, and a great game might have an initial retention of six months. (Best I've seen is 9 months for AC1, back when it had very little competition; worst is... well, not renewing after your free month is up.)

    After that initial period, players should hopefully go into a "rebound cycle" where they disappear for N months and then return to play some more for M months. This is, I think, the best-case scenario for an MMO and what it should shoot for.

    When a game damages the rebound cycle they shoot themselves in the foot. SWG's NGE did that for a lot of people. For me personally, DDO burned the bridge by making me quit in frustration too many times.

    On the other hand, most MMOs I play leave me wanting more fun, but finding only emptiness instead -- either lack of content or lack of friends or lack of compulsion to play. Those are all fixable things. In Lotro I simply couldn't find people to play with; now that it's "free", I intend to go check it out. They didn't burn their bridge.

    I think you can look at game-imbalance issues in this light: do they cause people to quit and never come back? Those are significantly worse than the ones where people quit because, even though they're kind of having fun, they stopped having people to play with.

    So the "leaderboard" problem you refer to is, I think, a lesser evil. The "slap you back to the respawn point because you picked the wrong power at level 3" problem is much more significant; both cause you to quit, but for one, you might come back some day.

    Deciding which balance problems to address is a very interesting and under-explored blog topic... though, I guess I may have just said about everything I have to say on the topic.

    Comment by Eric — 22 January, 2011 @ 1:31 AM

  18. Classes, Trinity and Balance, Oh, My

    [...] Psychochild: Stay Classy [...]

    Pingback by Tish Tosh Tesh — 22 January, 2011 @ 6:44 AM

  19. Dave Toulouse wrote:
    People play them like a leaderboard instead of trying to have fun....

    I think that's primarily because we cater so completely to the Achiever mindset. I think if we tried to appeal to another set, particularly Explorers, we'd see a different mindset. The problem is that Achievers are a well-understood group and we can cater to them easily.

    Eric wrote:
    I think you're wrong about the "false dichotomy" you discussed. Leaving aside terminology (which I agree is always difficult in these things) I don't see a continuum of choices: either you gate people from having multiple role-defining abilities or you don't.

    Okay, my confusion set in when you said in a comment previously, 'The "classes versus open skills" continuum is a large one, but you're making it sound linear.' I took that to mean that you agreed this wasn't a binary choice overall. But, I take it from this quote that you believe it does boil down to a binary choice. I have two thoughts.

    First, I don't think that having "multiple role-defining abilities" necessarily makes it orders of magnitude harder to balance a game. Yes, if you allow players completely free choice to cherry-pick the best abilities then you will have a degenerate system. This is why I wanted to discuss this issue in terms of "structure" where you can design how choices can be made, up to the point of restricting player input on character advancement to only being able to buy certain abilities at pre-defined levels.

    Second, the issue of "roles" is perhaps getting overlooked. I think the real divide here is about who should define roles. Your writing indicates that you think the designer should define roles precisely and players pick one. I'm more of the mind that the designer can define various options, but then allow players to fill the roles they desire. This might be why I like DDO currently, because D&D didn't exactly have the concept of the "holy trinity" of roles. Sure, in DDO you can build a "tank" with the Intimidate skill and Clerics have healing spells, but their defined "role" isn't simply to maintain aggro and keep casting spells to keep health bars topped off as it was in other games. So, you have the flexibility to pick a role you want without being pigeonholed into a specific role based on class and/or spec.

    But it's true that a PvP game that focuses on small-numbers encounters (typically 1v1) like AC1 (and Meridian 59?) is the best-case scenario for allowing multiple roles at once.

    Depends on what you call "small". Most of the meaningful fights in M59 were guild-on-guild, with a half-dozen people on each side and a few others as innocent bystanders or meddlers looking for a quick kill. One-on-one did happen, though, but it was mostly of the "gank and run" style. My goal for balancing was in the larger guild fights, since one-on-one fights were almost never going to be "fair" fights on even footing even ignoring character makeup.

    To apply my "classes" terminology practically, AC2 is a great example of a "classes" game, despite having no official classes. You certainly didn't pick any class when you started the game. Instead, your race choice locked you into a dozen skill trees. You could pick from many of these at whim, which provided you with a great many utilities and secondary attacks, but you could only get into one of the six specialty trees available to your race.

    In my continuum, I'd also see that as a fairly structured system. Your primary abilities come from a set of exclusive collections of skills, similar to classes; I assume you could have probably called those six "specialty trees" classes and not have surprised many people. The other trees add a bit of flavor to a character and customization, but don't define your character as much.

    (I'm referring to Combo Points vs. Rage versus Mana etc.)

    Funny you bring that up as a complicating factor, because I was thinking this is one area where a more structured/class-based game shines in being able to introduce different kinds of resources easier. In a less structured "skill-based" system, it'd be hard to introduce different resources like this without adding a lot of work, I think.

    Some more thoughts. Thanks for the interesting discussion.

    Comment by Psychochild — 22 January, 2011 @ 3:54 PM

  20. Best of the Week — 17-23 Jan ’11

    [...] no no‘s I can see that he does have some points. Psychochild takes up the gauntlet and argues for open skill systems, taking the issue from that of a binary Classes vs Open to More Structured and Less Structured on a [...]

    Pingback by Fun in Games — 22 January, 2011 @ 5:34 PM

  21. Perhaps part of the problem is that while in tabletop gaming, you can have tons of fun outside of combat - very few MMOs have figured out how to make group activities besides combat "work". You can craft, sure, but it's a solo activity (minus pyramid building parties in A Tale in the Desert). This makes it harder to define meaningful roles for people to fill. Combat comes to like four roles: damaging, healing, taking damage, or fiddling with buffs/debuffs. If you mix up the first three of those roles too much, people become too good at the one meaningful activity in the game (combat) too quickly.

    One "first step" in the right direction might, maybe to be replace drops with treasure chests and make quests more about getting to that chest, regardless of how you do it. Now you're open to tactics involving diversion, stunning, cherry-picking targets, and sneaking - increasing the roles available and thus making it "easier" to allow mixing of roles.

    Disclaimer: I have no MMO development experience. I could be way off on this - maybe what I just laid out quickly degenerates to having two roles: someone to grab aggro and run away, and someone to run around them, and the other roles might be there but useless. Maybe. My point, though, is that games have to be something besides "all combat all the time" if they want to expand the roles meaningfully. This is significantly easier in single player games.

    Comment by silver — 22 January, 2011 @ 10:06 PM

  22. silver wrote:

    "first step" in the right direction might, maybe to be replace drops with treasure chests and make quests more about getting to that chest, regardless of how you do it. Now you're open to tactics involving diversion, stunning, cherry-picking targets, and sneaking - increasing the roles available and thus making it "easier" to allow mixing of roles.

    Funny enough, that's what DDO does for the most part; although they give xp bonuses for killing more enemies, smashing items, etc., so it does kind of still boil down to slaughtering as many enemies as you can. But, the focus on having chests instead of lootable bodies does seem to speed up the game. As I've mentioned above, DDO doesn't follow the "holy trinity" slavishly, although certain skill sets do make the game noticeably easier in some situations. I think that's partially because the original pen and paper game was fairly flexible, but I think Turbine could have done more to encourage this through level design. Eric will probably say it would have been too hard to balance if they had done it differently. :)

    Comment by Psychochild — 23 January, 2011 @ 12:16 PM

  23. One problem is, I have this one good idea, but no follow up ideas to make a whole MMO game that lasts a long time for a lot of people, but I might as well share my one good idea.

    So, you go to castle XYZ and meet with the leader who tells you "all our forces are involved on the western front, but there's a rumor of a couple bands of BadDudesLikeOrcsOrSomething to the east who are getting together in some caves. If they do are really doing that, and really come after us, it would be very bad for us to try to fight on two fronts. Go Deal With It."

    Now you can go east and find the BDLOOS camps - and sure enough, they do seem to be amassing a force together. You have options to deal with them:
    1) under a white flag talk to one of the leaders and convince him the other one is going to backstab him. fake evidence for bonuses.
    2) sneak into one of the camps and plant evidence of the same, watch them fight each other.
    3) outright slaughter them
    4) attack some of them but run into the other camp and try to get them to mix it up with each other before they notice you were there. disguises are good.
    5) collapse the cave entrances at night. deal with the night watchmen. problem solved.
    etc. etc.

    I think I pine too much for pen and paper games, and MMOs aren't filling the gap. Maybe someday.

    Comment by silver — 24 January, 2011 @ 1:25 AM

  24. You Gotta Have Class… Or do you?

    [...] Psychochild identifies the untruth in assuming that the two ideas — skill based and class based — are exclusionary. He would change the terminology used to that of one on a continuum, with classical sense of classes being at the ‘Most Structured’ end, and completely open skill systems (like Asheron’s Call, Ultima Online, etc) being at the ‘Less Structured’ end. [...]

    Pingback by Fun in Games — 24 January, 2011 @ 6:57 AM

  25. Eric posted a followup article, but didn't link any of the discussions to create trackbacks.

    http://www.eldergame.com/2011/01/more-on-classed-vs-unclassed-games/

    Comment by Psychochild — 24 January, 2011 @ 12:10 PM

  26. So I took you at your word that you don’t post much and haven’t checked your blog recently. But this post reminded me of something that I’ve never understood about this argument so I’m replying anyway :)

    @Eric said:

    “…as long as you could be a full-power healer plus a full-power blaster…”

    I guess I don’t see how the above could exist. There is a saying: “Jack of all trades, master of none.” It seems to me that if you have 2 characters with the same “number of points” (whatever measurement system you use for increasing character growth), if one character puts all of those points into healing and one character divides all his points between healing and blasting, the first character MUST be better at healing than the second character. I understand the desire for people to be a snowflake – to want to be different from everyone else – but if the best healer is the person who takes the most healing skills…

    I also don’t understand the aversion to “class”. I suppose that if no one is a “Tank”, you can still handle mobs and be successful. But in a static group (Guild, raid group, etc.) its useful to understand who will perform which roles rather than try to manage things “on the fly” in every encounter. And how would you deal with a PUG where no one had a role? I suppose you could create a game where the power level of the characters was immense vs. the power level of the mobs such that even someone who divided their skills among several roles could still kill the mob and either not need healing or is able to heal himself. But I thought that “serious gamers” would stamp that “easy mode”.

    Comment by Djinn — 31 January, 2011 @ 1:32 PM

  27. @Djinn: there are a few ways to get a full-power healer/blaster in a system, and it's easier to have happen in skill-based system because of the flexibility that is one of the goals of that kind of design.

    For example, in UO, there was a hard limit of 700 skill points total, and each skill had a max possible score of 100. Problem was, to be a top-flight warrior only required 3-4 skills, and a top-flight mage only required 2-3. (I believe the idea was that players would have a focus ("choose a role") in combat, and then use the other craft/harvest skills to season and flesh out their characters... players had other ideas.)

    I don't recall the details on AC1, it's been too long, but a similar situation existed there, if I recall.

    Other possibiliities include: a lapse in skill/ability design, such that the most useful powers are not actually at the top of the talent tree/don't require all the available points; no hard cap on the number of skills/points/what-have-you that can be earned over the life of the character (no "max level"), combined with an (effective) maximum advancement in each skill/role.

    -=-

    I can't speak for others, but my aversion to classes is a matter of multiple issues, and has evolved over time.

    1) many role(s) I'm truly interested in playing are never valid in the game. I had a thief/mage back in AD+D that I'd love to recreate... even DDO was largely a failure on that (hint: even in AD+D, he wasn't really combat-oriented). From Rolemaster, I had a Mystic ship's captain/merchant, a pacifist, who got his first "kill" XP at level 12. Even something as minor as a ranger that is a master of spears, not swords and bows, ends up being either impossible or wildly gimped, because it didn't fit the designers' "vision" for the class.

    2) I've lost count of how many times I have logged off in disgust after an hour of "standing around" because "we need a tank... but none are available". Or spend 3 hours penetrating an instance and realize that "the final encounter really _does_ require an AoE crowd control, doesn't it... well, _that_ was a waste of time". Or, "The healer's kid is sick, he can't make it. Well, guess we may as well log for the night..." Not really the designer's responsibliity to force players to play the assigned classes/roles in specific proportions, to be sure, but it hasn't exactly endeared the design element to me, either.

    3) the roles that are made available, I've done to death (my opinion). I have 1000s of hours of MMO play under my belt over the past decade plus, including raiding, PvP, guild leadership, and so on. I sincerely doubt anyone can come up with a "Tank", "Healer", or "DPS" activity that I haven't already done to death. I get about 20-30 hours of play out of each new MMO, including alt-ing thru all the different roles, before I set it aside and sigh "I've already played this game." And that's been the case for about 3 years now.

    None of these are problems necessarily specific to classes, and none of them are without workarounds and ways to mitigate... but to my knowledge, they haven't been. (yet?)

    Comment by DamianoV — 1 February, 2011 @ 6:04 AM

  28. @DamianoV Thanks for the input. I guess the biggest issue in both questions is the style of MMO we're talking about. I was particularly thinking of your comments about Roles and not being able to play what you want. If the MMO revolves around killing 10 rats, at base your character needs to be able to kill to some extent. If you only run with a group you don't necessarily need to kill, but you need some kind of support role (healing, traps). I can see where it would be difficult to play a pacifist or a character that doesn't fill a particular role.

    But it seems to me that the issue isn't "I can't make my character the way I want" so much as "there isn't a game available where the character I want to make will fit in".

    Comment by Djinn — 1 February, 2011 @ 9:54 AM

  29. Djinn wrote:
    So I took you at your word that you don’t post much and haven’t checked your blog recently.

    Yeah, yeah, I've just been more inspired than normal lately.

    If you only run with a group you don't necessarily need to kill, but you need some kind of support role (healing, traps). I can see where it would be difficult to play a pacifist or a character that doesn't fill a particular role.

    For me, the issue comes down to what the player's participation is. In WoW (and others), the game and the roles are defined by the developers. You can't (currently) play a healing Warrior, for example. Not to say that WoW couldn't add some special abilities to allow Warriors to be healers, but it requires the developers to add support for it.

    The other end of the spectrum is where you allow players to define their roles. Instead of plotting out the whole game, it's the responsibility of the developer to set up challenges for the players to tackle with their characters. The player gets to pick what skill sets they want to use while playing, within the limitations set by the designer.

    Of course, this means that design is "harder" if you allow players to define their own characters. You can no longer create a perfectly scripted encounter (like some dungeon encounters most WoW raid bosses) because you don't know what players will bring and you can't even suggest what they should bring.

    This doesn't mean that class-based systems don't allow any flexibility at all. As I've mentioned before, I took a spec that allowed me to tank, DPS, or do a bit of healing in a pinch. One of my favorite memories in EQ2 was when my Necromancer, who specced a lot into the meager healing abilities, was able to heal the team the last little bit when the primary healer ran out of mana; I almost certainly saved us from a wipe after a bad pull. Likewise, people can create unimaginative characters in a less structured game and not have the flexibility they desire.

    Ultimately, I'd like to give more power to the players to play the type of characters they want. Then give them interesting challenges to solve using those characters.

    My further thoughts,

    Comment by Psychochild — 1 February, 2011 @ 11:18 PM

  30. One thing about classes that isn't mentioned often is that classes are lore. A Shadowknight class implies a lot of things about the game world. I think that could be leveraged more, for example if game races have different gods then maybe the priest class should be at least partially different for each race - same template but with some specific differences.

    Secondly i think classes should be used to match play-styles (or even roleplay styles) and not just group roles. For example a ranger could be designed primarily for outdoor solo (and secondary outdoor group utility) rather than DPS. A rogue for indoor solo (and indoor group utility) rather than DPS. The standard warrior, cleric, mage class could be designed for groups with an exception class (e.g barbarian, druid, necromancer) designed for solo.

    This would be explicit and no complaints after from solo classes wanting to be group-friendly or vice veras.

    Comment by tupodawg — 4 March, 2011 @ 2:29 AM

  31. Vocational Guidance Counsellor

    [...] while, but less structured skill systems can be a pretty tricky proposition as per the big debate a couple of months back. A model that would seem to fit rather nicely would be the tiers of increased [...]

    Pingback by Killed in a Smiling Accident — 9 March, 2011 @ 12:08 AM

  32. All the ills of mankind, all the tragic misfortunes that fill the history books, all the political blunders, all the failures of the great leaders have arisen merely from a lack of skill at dancing

    [...] vs open skills is always fertile ground for debate, and reaching all of level three is no position to render judgement on the system of Fallen Earth. [...]

    Pingback by Killed in a Smiling Accident. — 9 May, 2012 @ 5:38 AM

Leave a comment

I value your comment and think the discussions are the best part of this blog. However, there's this scourge called comment spam, so I choose to moderate comments rather than giving filthy spammers any advantage.

If this is your first comment, it will be held for moderation and therefore will not show up immediately. I will approve your comment when I can, usually within a day. Comments should eventually be approved if not spam. If your comment doesn't show up and it wasn't spam, send me an email as the spam catchers might have caught it by accident.

Line and paragraph breaks automatic, HTML allowed: <a href="" title="" rel=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <div align=""> <em> <font color="" size="" face=""> <i> <li> <ol> <strike> <strong> <sub> <sup> <ul>

Email Subscription

Get posts by email:


Recent Comments

Categories

Search the Blog

Calendar

September 2014
S M T W T F S
« May    
 123456
78910111213
14151617181920
21222324252627
282930  

Meta

Archives

Standard Disclaimer

I speak only for myself, not for any company.

My Book





Information

Around the Internet

Game and Online Developers

Game News Sites

Game Ranters and Discussion

Help for Businesses

Other Fun Stuff

Quiet (aka Dead) Sites

Posts Copyright Brian Green, aka Psychochild. Comments belong to their authors.

Google