I'm basing the idea of anti-patterns on the excellent post from League of Legends designer Zileas (http://na.leagueoflegends.com/board/showthread.php?t=293417). An anti-pattern is not a "you should never do this" rule. Rather, it is something that has generally will have ill effects on your game but WHICH MAY BE WORTH IT. Thus, when you see yourself following an anti-pattern, it doesn't mean that you need to immediately change things. Rather, you should take a deeper look at that aspect of your game and consider whether the thing you're creating is really worth the high cost or if there is a better approach.
There are several things I consider to be anti-patterns that are often used in MMOs without significant benefit.
1. Making interesting items, then overshadowing them with stat-sticks:
WoW offers an easy example of this by looking at early-game trinkets vs. late game ones. A level 40 character is going to be filling up his trinket slots with things like a gnomish battle chicken or a shrink ray: Interesting, fun, and cool items of varying degrees of value. A level 70+ character will fill all his trinket slots with yet another +X strength or +X intellect.
Simply put, they went through the work to make the cool stuff but then made it inferior to items that simply make your numbers higher.
A better Alternative:
The obvious answer is simply to make the interesting items more powerful, but there are a couple other details. First, expose the details (i.e. don't just say "Summons a crab to attack your foes". Give them an idea of how powerful this crab is). The other is to make sure you consider the comlexities of difficult content and the chaos of groups. If your armor has a chance to put the attacker to sleep, but a random AoE from the four other players will instantly wake him up every time (or if has no effect on the powerful characters where it would be most useful), then it's not a very good item design.
2. Awesome loot is limited only to certain levels or enemies:
If you go back and play many semi-old RPGs with randomized loot (Diablo, Might and Magic, Ogre Battle), in any given fight you have a chance for a really amazing drop that you can use right away. The general approach with MMOs seems to be that awesome items are only dropped by top-level dungeon bosses.
A Better Alternative:
Every enemy in the game (the only possible exception being the early enemies where the player is still learning the basics) should have a chance to drop something amazing and that chance should be high enough that players can expect to get something now and then. With this, there is a degree of excitement the entire way through the game, and even repeating content will feel less like a grind because there's always a chance for something great.
3. Creating variety by suspending game mechanics and character traits:
We all love variety, but there are better and worse ways of creating it. One that I have seen all over in RPGs, including MMOs, is where some task replaces everything about your character with a different (and generally far less interesting) set of rules. Whether it's the defense scenario in Final Fantasy VII where you were essentially playing a weak RTS or the Guild Wars 2 Claypool heart, which is filled in through going to combat instructors and doing what they tell you (which, besides being tedious, has nothing to do with how combat actually works in the game).
Sure, mini-game type things can be fun (although I don't believe they should generally be used to gate non-mini-game content), but it doesn't mean that there isn't usually a better way, as described below.
A Better Alternative:
Twist the rules without abandoning them. Instead of creating a new game that will take .1% of your development efforts, utilize the system that you poured months or years into perfecting but just in a different way. To use Guild Wars 2 as a positive example, Scenarios often require you to use a special item which will replace your weapon skills, but it leaves your utility, healing, and elite skills, as well as your stats and equipment intact. Consider unusual enemy stats or spawning rules or an escort quest where the person you're protecting actually has cool abilities that integrate with the player's.
As with the other anti-patterns, this isn't an exact rule for all situations, but if you're designing the "gather up the rabbits before they eat the watermelons" event in Guild Wars 2, see if you can come up with a way to make the existing class abilities affect how they accomplish the task (yeah, I know that's a hard one, but it's worth some thought).
4. Rewarding players for easy gameplay while punishing them for challenges.
Option 1: Go into a dungeon below level and 1 person short. Go against incredibly difficult challenges which require high skill and creativity. Result: After a long series of fights and a high repair bill, you get a bit of experience and some items that you can't use until you level a couple more times.
Option 2: Go into a dungeon a little above the intended level with a full team. Crash it effortlessly and quickly getting far more experience for the time spent and loading up on items you can actually use.
The response I most often hear when this is brought up is "Well, if people truly enjoy the challenge, then they shouldn't need rewarded for it." This view shows an extreme lack of understanding about players. Look at any online game for a few minutes and see how many people use whichever tactic is currently considered overpowered and how many use what is considered weak. The obvious message is that players will choose what works. Even though they play games to have fun, they're focus while playing is most often on what will accomplish their goals most efficiently (and in an MMO, those goals usually center around getting levels and loot).
Simply put, if you give players an easy, boring path to their goals and a difficult, interesting one, most will choose the easy path then complain that your game is boring.
A Better Alternative:
Talk to MMO players, and the majority give the opinion that leveling is a bit of a chore while end game is where things get interesting. The look at the games themselves and notice how the leveling player gets the best rewards for taking the path of least resistance while the max-level player gets the best rewards for challenging themselves. It can also quickly be observed that most players follow the incentives.
What this suggests, then, is to change the loot and experience algorithms to incentivize challenge. Dramatically increase the probability of good drops when dungeons or other events are undermanned and underleveled. If you want to take it further, make special events or such that require it (ex: A side boss in a dungeon only shows up if every character is under X level).
In listing my anti-patterns, I tried to hold to things that could be reasonably implemented into most existing MMOs whether in current content or future. There is a lot of room to grow even within the current formula, and I believe that analyzing against these anti-patterns (and others. I may make a part 2) would be one excellent way to gain that.