The party economy
Ok, enough background. If we think about our character classes as characters, our mental models are going to be shaped and limited by other games and other forms of fiction such as books and movies. Authors don't have to worry about Robin Hood being overpowered because he has high levels of both ranged and melee DPS, but game designers do.
Instead, I'd like to propose a much less warm and fuzzy way of analyzing the give-and-take between classes. The natural psychological forces pushing players towards tank-mages and character specialization are relentless, and the barriers we put up to stop them must be equally unromantic and purposeful.
What I'd like to propose is a simple framework for connecting MMO design to economic theories of human behavior, with the intent of providing some useful tools for thinking about class design. I am most assuredly not saying that everything about games can be described in economic terms, or that economic theory can be blindly applied to game design.
What I am saying is that there are some clear analogies to be made between class/party dynamics and some basic economic theories, and that a thoughtful exploration of these analogies produces a useful framework for thinking about class design plus a number of practical recommendations.
The fundamental bridge between these two areas is the idea that any group of characters, from a duo to a hundred-player raid, can be thought of as an economy with the characters serving as both producers and consumers.
There are a variety of "goods" or "products" that the characters create and consume, such as healing, damage, tanking, buffing/debuffing, crowd control, and utility abilities like stealth and summoning. These goods have costs, either explicitly (casting a spell requiring mana) or implicitly (opportunity costs, where every moment you spend casting a healing spell means you're not casting a damaging spell).
Every time a character benefits from what another character does, they're participating in trade with that character. When a tank holds the dragon's attention so that it doesn't attack a teammate, the tank is trading their production of tanky-ness to the other character.
Presumably the other character is producing and trading something the tank needs, such as sufficient healing to keep the tank alive or enough DPS to kill the dragon. Together, the group produces enough of each key commodity (tanking, dps, and healing) to complete their goals as a group.
Absolute and Comparative Advantage
The principle of absolute advantage was developed by Adam Smith in the late 18th century to describe some the effects of international trade. Smith wrote:
"If a foreign country can supply us with a commodity cheaper than we ourselves can make it, better buy it of them with some part of the produce of our own industry, employed in a way in which we have some advantage." (Wealth of Nations, Book 4, Chapter 2)
This is a clear argument for specialization, and the analogy to MMO classes is obvious. In a situation where each character has a role where they're the best (an absolute advantage), they should specialize in that role. However, hybrids can never be the best at any role, or they would replace their parent specialist class completely.
In that situation, we turn to a closely related theory called "Comparative Advantage". Attributed to David Ricardo and Richard Torrens, this theory states that each entity (countries in their examples, classes/characters in ours) has an area of production they're "least worst" at -- their area of "comparative advantage" in comparison to the other entities in the economy.
In our party economy model, consider the following ten character party makeup:
In this example, the hybrid character has a comparative advantage as a healer. The specialist DPS characters can produce DPS more efficiently than the hybrid can, and the specialist healing character can heal more efficiently than the hybrid can.
However, since there are lots of DPS characters present, the hybrid is going to be pushed towards healing, an area in which he has a comparative advantage over most of the party. He doesn't have an absolute advantage (the healing specialist does), but the presence of so many other DPS characters in the party economy means that his greatest contribution will be as a second healer rather than as an eighth DPS character. He has a comparative advantage as a healer and a comparative disadvantage as DPS.
The interesting thing here is that the hybrid has a comparative advantage as a healer even if the hybrid is a more efficient/effective producer of DPS than the specialist DPS characters. The strengths of a character in isolation (their soloing abilities) are irrelevant; it's their relative value in a particular group composition that drives the character towards one role or another.
Furthermore, if that hybrid character is always pushed towards healing, their player will begin to wonder whether they should have just chosen a dedicated healing class in the first place.