Fourth edition Dungeons & Dragons has adopted the MMORPG idea that character classes are specialized to perform certain roles in the party, namely defender, leader, striker, and controller. I've been pondering my impressions of how this works out.
I'm comparing this in my mind to traditional role-playing games of the past. In particular, Champions comes to mind is a great tactical role-playing game, since the idea of roles seems to be a rather tactical idea. In Champions, it is traditional to describe characters as being “bricks”, “martial artists”, “mentalists”, and so on, each of which is assumed to have certain characteristics. But these are just labels - there is nothing that constrains you to make characters that fit any of these niches, and even if you do, the roles don't really dictate what your actual purposes in a fight, they are more like descriptions of what you tend to be good and bad at, from which your purpose might be inferred. In Dungeons & Dragons, each class has strong specific powers to perform its specific role, and there is a clear idea of how the different roles are supposed to work together in combat.
What I think I like most about the roles is the idea of the leader role, a character who specializes in making the party work better as a whole rather than in making individual attacks. The main reason this appeals to me is that it allows certain character conceptions to work much better. I've created a number of character conceptions in the past which are based on the idea of being leaders or support characters. Certainly in fiction, the idea that someone can be a great leader is considered very important and meaningful. But in game terms, this typically had no effect, or very little effect. And trying to be a character who offer support in other ways was typically boring, ineffective, or not well supported by the game system. The leader classes allow you to make this character type interesting, effective, and fun to play.
The defender role is also fairly interesting and had a lot of potential. In a game like Champions, a character with high defenses would often try to suck up attacks by going one-on-one with the most powerful opponents were standing in the middle of the toughest situations, but there wasn't really anything in the game rules that specifically made this work. Adding in special abilities to make this work as part of the game is fun because it allows you to better implement the purpose of your character, and having more types of useful abilities makes characters more distinctive and thus more interesting.
In 4th edition Dungeons & Dragons, not only do they create these new powers to represent the various roles, but they make sure to specifically give each class powers that strongly place them in a given role, so that each character is a specialist in one role, rather than having a general assortment of various powers. The idea is that the party forms an interlocking whole, with each character participating in their own way (as in World of Warcraft, which I've heard about but never played).
So from a game design perspective, the question that comes to my mind is whether specializing each player into a specific role works better than my previously preferred character generation style of giving each character whatever powers seem appropriate for their conception, even if the powers do not synergize together to implement a specific role. After playing Dungeons & Dragons for a while, I feel that having a party which consists of an interlocking group of characters, each of which sharply implements one of the 4 roles, works well but isn't particularly any more (or less) interesting or desirable than a group of characters with fuzzy roles.
First of all, each of the 4 roles seems optional and the party is still interesting without each one. While having a defender in the party is certainly useful, the opportunity attack mechanic seems to do a pretty good job of tying down monsters in melee combat with the characters, and the defenders mostly just tie down the same monsters even better. Just like in any other game, the way the damage gets spread out among the players ultimately depends on who the GM decides to attack. The striker role is really just better at the “default role” of all characters, causing damage to single targets, so removing all strikers wouldn't really change much. The controller certainly has a big impact on how effective the party is vs. large groups of monsters, but I've played with and without a controller, and although it changes the feel of combats vs. masses of monsters, the combats are just as fun without a controller as with a controller. The leader class is the only one that is absolutely essential, but that is just because the existence of healing powers in the party makes the damage system work much better. If you look at the “leader” aspect of the role, the ability to give bonuses to other characters in the party, this is interesting and certainly useful but the combats seem like they would also work okay without a character dedicated to these powers.
It is probably more interesting to have a variety of the different types of powers (certainly I think it wouldn't be that great if everyone was just a damage dealer and no one in the party had tactical powers), but it doesn't seem essential that these powers be given to characters of the corresponding role.
The downside of strongly identifying each class with a specific role is that can be a definite nuisance when trying to create exactly the character you want to create, because it forces the character to have or not have abilities in a way that may not match the "role-playing meaning" of the class. The problem is subtle, but quite noticeable to my perfectionist eye.
For instance, the rogue class has a very powerful sneak attack ability to fulfill its role as a striker. However, the "color" of the class makes it appear that a rogue can be a wide variety of dextrous, skillful, lightly equipped martial character types. You can be thief, a swashbuckler, a ninja, a bold explorer, or lots of other things, and the powers and paragon classes seem to support this. But ultimately, the characters you can actually create are more limited. You can be a thief who is obsessed with gaining combat advantage, a swashbuckler who is obsessed with gaining combat advantage, a ninja who is obsessed with gaining combat advantage, a bold explorer who is obsessed with gaining combat advantage, and so on. The idea of playing a character who is obsessed with gaining combat advantage is pretty interesting, but it is a little monotonous that all rogues must be built this way.
A similar example is the fighter. It sounds like the most basic fantasy archetype you can have, just a heavily armored martial warrior. But ultimately, you can't just make a heavily armored mercenary. You have to be a heavily armored mercenary who immobilizes his opponents. This is a cool idea, but it is odd that every fighter has it. What I've seen is that when players who aren't too familiar with the rules make a fighter, they tend to envision that they are making a heavily armored warrior with a strong defense and powerful offense. The idea that the character class has this very strong ability to immobilize opponents is just kind of weird, they aren’t necessarily all that interested in having this ability as compared to something more straightforward.
A final example is the warlord. The idea of the great leader who gives benefits to his troops is really cool. But from a role-playing perspective, which you usually envision is a great warrior with strong leadership abilities. Having every single power, without exception, be based on teamwork and leadership, can feel somewhat restrictive.
I think there is actually a lot to be said for making sure that the characters have distinctive powers and capabilities, so that each character adds something to the party which no one else does, and the characters do not overlap. However, I do not think sharp role delineation is necessary to achieve this - in fact, it seems to get in the way (characters of the same role tend to be pushed towards more similarity with each other than they might otherwise prefer).
Actually, now that I think of it, the striker role is so generic that my issue with the rogue can't really be with the role, but more about the very specific implementation of striker power that they have. If the rogue was simply allowed to pick a different striker power, you could probably make a swashbuckling fencer just fine. But what if you wanted to make a fencer who could sometimes attack fiercely, sometimes inspire the party with his charismatic leadership, and sometimes tie down the opponent with a defensive fighting style? This is the sort of thing that I think would work just fine if allowed, even though it is spread out over multiple roles.