“Level Design is the data entry and layout portion of the game development cycle. A level is, for all intents and purposes, the same as a mission, stage, map or other venue of player interaction.” - Tim Ryan, Gamasutra
From the title screen onwards, nothing should remind them they are in the real world. The player must attain a level of escapism for immersion to take effect, which is only achieved by making sure nothing reminds them that they are sat at a computer, jamming with a keyboard.
Immersion can be broken by a variety of common mistakes: Graphical glitches, bugs.
In terms of design, something like a car going over a bridge in the background of a medieval setting. Things like being told to 'press R to reload' or 'right-click while pressing W' can break immersion as it reminds the consumer that they are playing a game.
Though some would not admit, we all love a challenge, and games deliver us these challenges frequently. Though there is indeed the overarching challenge of merely beating the game there are also mini, sub challenges throughout most games that constantly keep a player on their toes like; Do x in y amount of time, or get from a to b while y does z etc.
Challenges within levels can range from traversing a hedge to surviving a room full of elite soldiers complete with rocket launchers.
The tough part with making a game challenging is not making it too challenging so that the majority find it more frustrating than fun. Emphasis on 'majority' here, as that is your main audience thus you need to find that medium difficulty that does not neglect most player only leaving the most skilful satisfied. This can be seen as unfair towards people of higher (or indeed, lower) capability, however not all is lost as games frequently have extra content to challenge skilled players and those who perhaps have already completed the game, and maybe might like to try something a little harder.
To clarify somewhat; if your game is a FPS, you should challenge your players aiming skills, from how accurate to how fast they are. If a real-time strategy game, challenge the players ability to manage his or her defence and offence while maintaining a steady economy within their base.
Challenges are essentially training, so its a good idea to constantly challenge your player as to get them up to snuff as games generally get harder as they go on.
Games, as do films and books and any other medium within the entertainment industry need to be... entertaining. If a player gets bored they could well go do something else. As a game designer this is death, you want people to play this game so for one to just turn it off should feel like a knife in the heart, and for the most part (psychological issues aside) it is the designers fault.
This reminds me of a game called Mass Effect, you may have heard of it. I did, and I also heard all the hype behind it, about how it is such a great story, completely immersing etc. So I finally bought it and began playing. Didn't think it bad by any means, in fact, I'd go as far to say I was rather enjoying it. Then the Citadel part came. Just one hour earlier I was fighting robots with my laser gun, throwing grenades at some resurrected blue buggers and unravelling a tale of struggle, impending doom and treachery but now... I am running around for three fucking hours on this massive ass 'town' going through loading screen after loading screen and talking to people I really could not give a damn about who have this problem or that problem and could I perhaps find it in my heart to do this or that for them and... you know what? I quit.
Assuredly I have a personal preference for combat unending over prolonged amounts of banter yet still this caused me to stop playing. Begrudgingly I picked the game back up a month or two later with the promise that it got better after the Citadel, and it did. Yet that doesn't forgive the game for it's massive flaw. Barely two hours into the game and I want to quit and let me tell you, I love narrative, I love story but do I give a shit about Joe Predicament? NO!
Robots are attacking human colonies and one of the Citadel's finest agents has betrayed us and I've had a vision of some evil big bastards coming to gobble up our galaxy so I'm now on my way to talk to the high council on what our next move will be, because it could well determine the fate of not just humans, but all other life in the cosmos... no, I'm not going to hand over your love letter.
Back on track.
We've talked about games needing to be immersive, challenging and entertaining. These are pretty much the building blocks a game is made from. A little further reading however.
A game should invigorate a players senses; when they go into an area their mind should race with ooh's and aah's, they should WANT to progress through the level, if they see something interesting, they will go to it. A nice trick with this technique for a complete level is to show the player where they will end up later. For example: looking out a window to the street below and then ending up down there in a few minutes, or showing a massive radiant light emitting structure a mile away only to end up there soon enough. Another good one is showing something like this throughout the level so the player not only can see where they are headed, and await it with anticipation, but also to give the player a sort of triangulation point so they know where they are in relation to where they've been, or possible in relation to somewhere else they have seen.
Other ways to subtley lead players through a level include lighting, or something as simple as where your enemies are. Don't know where to go? Follow the gun shots. Simple, subtle, works every time.
Players also require freedom. This is their game, they've bought it, it is theirs to do with what they want. Telling them they must go down this alley, although a simple command to follow, is somewhat condescending. Like the big arrow in the old Streets of Rage games telling you to go forward... yeah no shit buddy.
Presenting the player with options is always a great idea. All you need to do is set the playing field and just let them work it out, kind of like a monkey with a children’s puzzle box. You don't tell them how to put the shapes in the holes, you leave them to it, trying to ram a circle into the triangle slot doesn't work so they try something else. Not only does this add a lot of variety to the areas, it also gives the player a quick sense of accomplishment, which is very powerful motivator.
Giving a player the option of two routes that end up at the same destination is another great piece of level design, the power is in their hands. Do they take the stealthy route through the vents or the balls to the wall front door approach? It's their decision to make and makes them feel like they're playing the game the way they want to, you're not penalising a stealthy player by making him go through the door. It all adds to the immersion.
With all this said, however, the main thing to keep in mind is that you must make the player care. Story plays a big part in that motivation however if the levels are poor (Citadel from Mass Effect) they will stop caring about the game. Make them exciting, make them want to progress, to explore and reward them for doing so and you'll find them squeezing in that extra level before bedtime.