One of the central decisions we faced when developing dMetric was picking between class and skill based character advancement. As detailed in the primer, we eventually resolved to use classes – a choice that initially felt odd. Not only have classes fallen somewhat out of favour in the current independent game design zeitgeist, but when we initially set out to design dMetric we were adverse to the idea entirely.
So what happened? Half of the reason was simplicity. Classes provide logical units to organize skills and abilities under. Having a player pick a class takes less time than having the player research and select a set of skills on their own. This reduces the overhead invested in character creation and reduces the chance of players creating a worthless character with mismatched skills. dMetric is all about simplicity and speed of play, so classes were the perfect fit in this regard.
Secondly, balancing skill based character advancement is a nightmare. This difficulty lies not simply in the sense of keeping the power of one skill on par with another. When we were brainstorming different skill based systems, we found that certain problems chronically arose. This is not to say that skill based advancement is fundamentally flawed, but rather that it possesses certain shortcomings that make it undesirable in some situations – specifically our own.
One of the easiest ways for skill based advancement to break a game is to leave it completely unrestrained. This “free market” approach, where a player may level up any and all skills freely, seems the obvious choice to foster flexibility and creativity in character building. Ironically, such a system can actually operate antithetically to this goal. If, given enough time, any player can do anything, individual strengths and weakness disappear. Why wouldn’t a wizard learn to use a bow? Or a fighter learn to heal himself? Why wouldn’t a character simply become good at everything if the option is there?
In the early levels this is hardly an issue, but as characters advance towards the upper echelons of power individuality progressively dissolves. This system also spawns strange hybrid characters that shatter the diegesis of the game. A character that is good at everything is uninteresting and bends our conception of what an individual human (or elf, halfling, gnome, etc.) is capable of. The difficulty of the game is also trivialized, as no situation arises that the character is unable to tackle on his or her own. So why not simply limit advancement by capping how far skills can be raised in total? Simple enough, right?
Capping skill advancement brings with it the issue of “builds.” Anyone who has ever played World of Warcraft is familiar with this terminology. In WoW, every player is provided a fixed number of talent points to pour into talent trees that modify and improve their core abilities. Hypothetically, this allows for near infinite shades of character colourization. In reality a general consensus exists in the World of Warcraft community that there are only one or two effective ways to allocate these points for each class (pending patch-by-patch balance changes). These talent “builds” offer the mathematically highest damage output, healing output, damage mitigation, or whatever other statistic is most integral to that specific class’s success in a group.
While it could be argued that this is a problem endemic solely to MMORPGs due to their being online and highly competitive, this type of metagaming is a recurring element of pen and paper roleplaying and of gamer culture in general. Googling any class in Dungeons and Dragons or Pathfinder will yields dozens – if not hundreds – of websites detailing the perfect ways to allocate attributes, skills, and feats. This isn’t necessarily a bad thing. In the best case scenario, this mentality can simply nurture more standardized characters that – while more rigidly conceived – can each still fill unique rolls in a group. In the worst case scenario, builds can arise that are powerful and versatile enough to jeopardize the integrity of the game, making all other builds worthless by comparison. Self-healing warriors that can battle endlessly or wizards who can annihilate any opponent with a specific combination of spells are good examples.
It’s possible to create roadblocks to prevent players from taking this metagaming approach. Instituting logical limitations on how skills interact with one another can dismantle these uber builds fairly easily. These limitations can range from practical considerations (no sneaking or spell casting in heavy armour), to rules tethered to the logic of the campaign setting (a character can’t be both a necromancer and a healer), to mostly arbitrary decrees (two-handed weapons can’t be used to backstab). These limitations force more pragmatic character building and limit the power an individual character can achieve. They also encourage creativity in character generation as players feel less handicapped by picking more atypical or socially oriented skills.
Asides from resulting in unwieldy appendices of rules, this method of dividing specific skills indirectly promotes class-like structures. If a character cannot cast spells if they wield or a weapon or wear armor, for all practical purposes picking up a single spellcasting skill turns him or her into an archetypical wizard. The situation holds true in reverse: if a player wearing heavy armor can’t sneak or cast spells, by donning a single piece of plate mail they have accepted the mantle of a stereotypical warrior. For all intensive purposes, these characters have chosen classes. The flexibility of skill based advancement becomes an illusion.
The best roleplaying games are the ones that manage to walk the tightrope between deep character customization and reasonable character power. One of the best games I’ve encountered in this regard is Fallout 3. Each skill – whether it be energy weapons, lock picking, or speech – is useful enough to be desirable but not so powerful as to be considered essential. Progression through the game also accommodates almost any character build, whether the player wishes to be a marksman, a smooth talker, a bare knuckle brawler, or a rogue. All this with very few arbitrary limitations!
The pitfalls I’ve mentioned are something all game designers should think about during game development. Being aware of these missteps early on in the design process prevents frustration and head-banging over balance concerns later on.