A few posts ago, I expended a page of text pondering some of the costs of encounter and adventure "balance," which in my mind at least added up to a compelling case not to bother with it at all. Now, another venerable D&D trope that I think can (and probably should) be left by the wayside: the Balanced Adventuring Party.
The standard Balanced Adventuring Party consists of a fighter, a magic-user, a thief, and a cleric. If there are more than four characters, the extras are usually extra fighter or cleric types, and all are within a level or two of each other. I'm really not sure exactly how or when this idea got started. The monsters section of the Moldvay/Cook edition has entries for entire parties of a single class (Veterans, mediums, bandits, and acolytes, respectively.) The rules for generating NPC adventuring parties involve rolling dice to determine the number of members and the class and level of each. The name of the game, it seems, is randomness, not balance. There's no mention of balancing the party at all. If anything, it's openly thumbing its nose at balance.
To be fair, I also don't know of any rule set that officially enshrines the Balanced Adventuring Party as something to be striven for, but quite a few adventure modules explicitly suggest and encourage it in very strong terms. I'm definitely not knocking the general idea that a diverse assortment of talents can be a good thing, but it seems to me that there are a couple of artificial impositions on the game that push the notion of party balance from a potentially advantageous option to a sacred principle of adventuring.
The first of those things is the Balanced Adventure, especially those at the railroady end of the agency spectrum. The party is expected to overcome a specific series of scenes and encounters en route to a predetermined goal. Because there's little opportunity to circumvent or avoid encounters, or to go off in search of more appealing challenges, each encounter and the adventure as a whole must be balanced to provide challenges that are neither insurmountable nor boringly easy - a task that's considerably simpler if one assumes a certain ratio of competencies in the party. Different sorts of obstacles are included in the adventure to give each class something to do and a chance to shine - tough monsters for the fighters, locks and traps for the thief, barriers that can only be solved with magic for the magic-user, undead for the cleric, etc. Many, if not most, commercial adventure modules that I've read and/or run seem to be written with the Balanced Adventuring Party in mind, and deviating from that formula too far is likely to make things a lot harder on the party, the scenes tailored for absent classes becoming either impossible or much more taxing to the party's limited resources than expected, throwing the balance for the adventure's climax against them.
The other artificial imposition is niche protection, the idea that the usefulness of a class depends on its being able to do things that the other classes can't. One of the most oft-cited arguments for niche protection is that of the magic-user class allegedly stepping on the toes of the thief class with its Knock and Invisibility spells. (By the same token, of course, fighters must be arbitrarily barred from forcing or smashing locks if the thief's supposed bailiwick is to be preserved.) If the thief's lock-picking niche is to be protected, then it follows that a thief is an absolute necessity to any party that expects to find locked spaces that it wishes to access.
I've already dealt with encounter and adventure balance here. As for niche protection, I think it's an unnecessary convention based on the misconception that class defines what characters can do, instead of representing a style and set of methods for approaching challenges. A thief's skill is but one way to defeat a lock - often the most effective and advantageous way, but not the only one. Magic is a very convenient method of defeating many obstacles, but oftentimes a little old-fashioned wit or muscle power properly applied will serve as well.
Once we dispense with those two contrived bogeymen, there's no reason why a party that lacks one or more of the archetypal classes can't be a viable adventuring team.
A party of all fighters can solve problems with brute force and skill at arms. When the fighters face a challenge to which their strengths aren't ideally suited, they can improvise a solution or they can shrug and go off in search of something that suits them better. They are neither constrained by their class as to what obstacles they can try to overcome, nor bound to tackle challenges they feel ill-equipped to face. A party of magic-users might lean heavily on deceit, negotiating, and running away from monsters instead of physical combat, but they aren't barred from mixing it up whenever they like their odds. A party without a cleric has other means of healing themselves and dispatching undead monsters, and a group with no thief can still negotiate a warren of traps and locked doors if they choose.
The bottom line for me is that, like adventures and individual characters, adventuring parties are far more interesting and varied when freed from rigid formulas, when you let unexpected things occur. Players shouldn't feel obligated to play a particular class just because the group "needs" one or already has "enough" of another. Whether you end up with a Balanced Adventuring Party, or with six fighters, or with four magic-users, a thief, and a halfling, what's the problem? In fact, given the option, I'd rather run a game for one of the latter two parties.
I'm seriously trying to come up with some way to have my players roll up their next set of characters separately, without any knowledge of what the others have chosen to play, to see if that makes a difference in the composition of the party.