The worst thing a game can do is bore the player. The easiest way to bore a player is to be repetitive. Whether it be reusing a puzzle, using endless waves of the same enemy, or the same game mechanic for every fight, it gradually will annoy me until I stop playing.
The first apple was great. The second was okay. The third was tolerable. The fourth made me vomit.
At that point, I'm unlikely to buy other games by that developer. So here's a few ways for a few of my favorite genres to avoid being dull.
Real Time Strategy
RTS titles have actually come a long way in avoiding becoming stale. In part, this is due to a lack of new releases in recent years, with only Starcraft II coming to mind as a major title this year. However, a much larger reason is that developers have shifted from a focus on starting with a small base, gathering resources to make your base larger, training and upgrading an army, and then advancing to destroy the enemy base. If you screw up somehow it means that you're set back a good half hour or more of playtime as you have to redo all these steps to try again.
It's a lot easier to blow up a house than to build a new one.
To evade this endless cycle, developers have added more creative objectives, optional side objectives that can make it easier to continue, and more personality to the game's storyline. For an example of the latter, Warcraft III featured an engaging storyline where you played with a Prince through his attempt to seek vengeance for an attack on his people by the undead. He then became corrupted and turned to the dead himself. In the next campaign you played him returning to attack his own people. Starcraft II included a number of missions where you were required to defend civilian populations, rescue prisoners, and perform other tasks that generally made the campaign stay fresh throughout the whole experience. A number of other games allow the player to pick different areas to attack, giving some interesting choices to spice up gameplay. Any new RTS game would be well advised to include these features if they want their game to impress anyone.
Unfortunately, there are still some weaknesses. In some cases, the campaign missions are just too long. For instance, there's one in Starcraft II where you are required to destroy a settlement that has been infested by the Zerg.
Infesting things always involves tentacles!
That sounds well enough, but there's something like 400 buildings you have to destroy, most of which take a long time to blow up. Additionally, you can usually only destroy them during the day because at night the buildings spawn large numbers of enemies that move toward your base. These enemies aren't THAT much of a threat but they do make it difficult to continue your offensive during the night. Even worse, the enemies are all of one type, a bunch of zombie-like foes who shamble without any strategy or tactics toward your main base. As the planet shifts between day and night every five minutes, you're left with having to spend somewhere around 40 minutes destroying enemies and buildings that aren't particularly engaging or interesting.
The original Starcraft campaign featured a similar problem in that the game featured three races, Terran, Protoss, and Zerg, but frequently in each campaign you would only fight one or those for the vast majority of it. This would get somewhat boring fairly quickly as you would keep having to kill the same types of enemy units. In the Terran campaign, you only got to fight against the Protoss once in the entire 10 mission campaign! If you go to the trouble of making multiple races for your game, you should at least allow the player to fight against each of them a few times.
Multiplayer games tend to avoid the problem of repetition because humans are far superior to the AI in their ability to use creative strategies with the tools at their disposal.
Although playing online does pose some other problems.
This makes sure that the game is always a challenge. However, if certain strategies are far superior to all of the other ones, the game quickly devolves in depth to become a matter of execution as opposed to strategy. This is why it is important for developers to continue to support their game through patching after it has been released. The balance problems, where certain units or strategies are far too strong or too weak, need to be addressed to keep the game interesting. With support, an online community can exist for years to come. People still play Warcraft II online and it came out in 1995!
And still features realistic explosions!
First Person Shooters
I addressed this a bit in my article about creating challenging enemies, so I won't add too much more here. Basically, avoid making the game into a series of identical firefights. Add something to spice it up. Make enemies be more intelligent, force the player to use more strategy, add some interesting set-piece events. A great example of this was in Call of Duty: Modern Warfare, where for one mission you were shifted out of the action entirely to be the gunnery officer on an AC-130 gunship.
It's a nice change of pace to be able to point and click to blow up the enemy. Leaves a hand free for a beer.
Another instance in the sequel to Modern Warfare was when you had to use a shield to repel enemy gunfire while advancing through a prison against the guards. The enemy bullets make huge divots on the shield and a loud thump as they ricochet off. It really is quite terrifying at first.
It takes balls to want the enemy to shoot at you.
Role-playing games
One thing many RPGs do is reusing the same model of an enemy for much of the game. Sometimes they color them green or red or something to try and trick my brain, but it doesn't work.
No, sticking a couple arrows in its head doesn't make it a new enemy type.
I want to see a progression of different enemy types throughout the entire game. I also want them to use different abilities and attack types. It's much more fun to play against a group of enemies that uses magic, ranged attacks, and melee attacks instead of just one of those three.
The best example I can think of for a game series that has done this consistently well is the Final Fantasy series. Throughout the entire game you face an ever changing mix of enemy models with different creative abilities that you have to prepare for. Almost every boss fight against stronger enemies has some unique elements as well, such as a specific weakness or a powerful attack that you have to prepare for in order to survive.
I would also add that RPGs that choose to use a leveling system need to have choices available when your character gains levels. If you don't have any options as your character becomes more powerful, the game becomes dull much faster. Ideally, these will make your character somewhat unique so that the game can feel fresh each time you play it.
One game that did this well was Dragon Age: Origins. For instance, if you played as a warrior, you could choose to use two handed weapons to deal high damage. You could use a sword and shield to absorb damage for your group. You could also use two different one handed weapons to deal high damage. If you were a mage, you could pick between a number of different types of magic to specialize in healing and support magic for your party, dealing damage to the enemy, or other options. Finally, you could pick two of four specializations for each of the three classes, allowing you to have access to different abilities. This level of customization helped keep the game fresh throughout, as you could choose to make your character whatever you wanted to be.
I chose to make my character show off his pecs.
Massively Multiplayer Online Games
I'll preface this by saying that the only MMO I play is World of Warcraft. However, any MMO that features RPG elements needs to feature all of the above RPG features as well as a few more. The only big annoyance I have with WoW is when the drop rates for items for quests are abysmally low. For instance, if you have a quest that requires you to hack off heads to return to a NPC, you can almost guarantee that half the enemies you find will inexplicably not have a head.
Even when you can clearly see a head on the ground with your knife sticking out of it.
Simply put, if you're going to have a quest that requires items to be found from an enemy, have the drop rate be reasonably high. It should not take fifteen minutes to collect items. That's not fun.
Travel time is also a big nuisance in some cases. Many times, a quest would have you go out and kill an enemy, go back to town to turn in the quest, whereupon you would get a new quest to kill a different enemy at the place you just visited. Why didn't they just tell you to kill all of the guys there at the same time!? Fortunately, the new expansion, Cataclysm, has pretty much removed this problem, but it's important to make sure that it doesn't come back.
A final complaint is when you have quests that demand that you kill one kind of enemy, but that enemy is nowhere to be found. For example, if you have a quest to kill 8 Scourge Siege Engineers at this one location, you may find that out of all the monsters at that place there are only 1-2 Siege Engineers.
But don't worry, there's at least fifty mindless zombies to kill in the meantime.
This means you have to wait for them to respawn, which takes quite a while. This is silly design and shouldn't be done. There should always be enough enemies to fulfill the number you have to kill.