The MMORPG of the future is going to have some trouble.
Right now, massively multiplayer games are essentially really long RPGs. Players level through them, reach the end, and eagerly await an expansion. However, balancing an RPG is a science which has yet to be founded. This means that the game gets rebalanced every month or so, especially right after a new expansion. Moreover, just the act of releasing an expansion rebalances (or unbalances) the game.
You can't leave a game unbalanced, or it becomes a dominated strategic space. IE, not a game, but a puzzle. A solved puzzle.
The thing is, this isn't healthy.
First and foremost, it requires serious manpower. All this data-mining and re-balancing takes employees, each of which costs probably $100,000 a year... or more! Certainly not affordable for a niche game, and even a big-name game should think twice before just dropping that kind of cash. Especially since the balancing is usually pretty tenuous.
Second, it pisses off the players. Rebalancing will always please some people and piss off others - but you don't need the additional pleasure and the additional pain is not at the top of your list, either.
Third, the law is about to hit hard. With real money trading, it won't be long before game administrators find themselves getting sued by people with tens of thousands of dollars invested in the game - tens of thousands that were just cut in half by a rebalancing effort. Especially plausable in games where one player can conceivably upset the balance simply by investing tens of thousands of dollars.
What's the alternative?
No rebalancing. Sure, have add-ons. Whether player-generated or corporate-made, they're all but necessary to keep a game like this interesting. But build the game so it doesn't require balance.
Partition and counter-weight your gameplay elements.
Counter-weighting is simply making every feature have a counter-feature.
For example, every game with stealth should have an anti-stealth. The ability to detect stealth in every manner that stealth gives an advantage. If stealth lets you rob houses, houses need to have traps. If stealth lets you steal from players, players need to be able to invest in wards and watchers.
Is stealth too powerful? All that means is a rise in the stealth-related population. Stealthy people become more popular because stealth is more useful. Anti-stealth becomes more popular because stealth became more popular. Stealth stops becoming more popular. Equilibrium is reached, albeit with a larger stealth-related infrastructure than you might have originally planned.
If you release "The Hounds of Orgoth", which gives thieves a significant bonus if they do such-and-such, you won't be unbalancing the game. All thieves will flock to Orgoth, but when they emerge all that will follow them is a horde of anti-thieves, getting fat on protecting people from thieves. Maybe the anti-thieves are also thieves... which brings us to the next point.
Of course, to do this you need to design your game with it in mind from the ground up. You can't reasonably expect to paste this sort of fundamental feature into a game that's in development.
Because none of the gameplay elements can stand alone.
No matter how many thieves or anti-thieves there are, they cannot do anything very well without non-thieves. Or, at least, thieves who do other things. Thieves, no matter how good at stealth, are greatly enhanced by having wizards, warriors, blacksmiths, alchemists, and any other class around. Not only because they're easy to steal from, but because the gameplay advantages they offer are overwhelming. A thief with a sword is better off than a thief without one. A thief with an invisibility potion is better off than a thief without one.
Because of this, even if thief was the uberclass, thieves would be all but forced to find and hire non-thieves to boost them. "I'll protect you from other thieves if you make me invisibility potions," for example.
The same could be true no matter who the 'unbalanced' class is. If it turns out that potions are insanely powerful, everyone will want to be an alchemist... until the market is flooded with potions and nobody to buy them. Self-correcting, self-balancing.
Of course, this produces a vibrant, complex, and thoroughly daunting world. Newbies and casual players will feel lost and outgunned.
That's where partitioning comes in. But this post is already long enough.