Tutorial 2: A Theory of Level Design

From the previous tutorial, you should now know the basics of working with brushes and creating levels. Knowing what makes a level interesting is more difficult than knowing how to make a level, however. This bigger picture, the best practices of level design, is the focus of this tutorial.

Game Design, Level Design, and Level “Style”

What makes a level good will depend on the game. Think of the game you are designing the level for, all of the player’s abilities, the items, the non-playable characters. Think of the way you want all of those systems and elements to intersect, and think of what you want the player to feel. Tossing all of these elements into a box may allow for some interesting interactions, but the more elements there are, the less likely any interaction will be. This toy box of game elements also means the designers have little control over what the player experiences; players are left to create their own fun.

To understand this difference between game design and level design, let’s look at explosive barrels. When an explosive barrel takes damage, it explodes, which pushes nearby objects away while causing them damage. On their own, explosive barrels are interactively simple. However, an explosive barrel next to an enemy allows for a more interesting interaction. Instead of attacking the enemy, the player can now attack the barrel to damage multiple enemies near it. For an even more interesting interaction, we can place barrels within the damage radius of other barrels; now destroying one barrel causes a chain reaction. Emergent behavior! Unless a designer places the barrels in a level to allow for chain reactions, the interaction may as well not exist, as far as the player knows.

Some of a game’s possible mechanics are more interesting or thematically appropriate than others. The function of level design is to focus, intensify, and cull the game’s mechanics, so that the game is the best it can be.

Consider games where players control a character, and can run and jump from a third person camera. From these mechanics, you could create a fast precision platformer, or a slow atmospheric exploration game. The difference lies in the level design. Some of these possible games are more interesting than others, and it is the purpose of level design to focus and narrow the game’s possibility space to its most interesting form.

Because of the importance level design has in focusing the game’s mechanics, it is good to create a common style for the levels in your game. The mechanics described above may allow for precision platforming or atmospheric exploration, but it is bad for a game to switch between the styles without reason.

Now that we’ve considered how level and game design differ, let’s look at the fundamentals of level design.

Elements of Level Design

  • Movement constraints. (Where can the player move? Is there a sense of direction?)
  • Occlusion. (What can the player see? What is hidden? How does moving through the space open and close sightlines?)
  • Item Ecology. (This includes items, player and NPC starts. Does the player need to worry about their resources?)

Environment art has its own elements, which overlap with level design in the way decoration informs the player about a layout. However, the degree of this overlap depends on the game and its art style.


At its most basic, the movement possibility space defines the boundaries of the players’ movement. This is the path that connects rooms, and this is also what makes these rooms interesting.

Some kinds of level geometry direct the player more than others. For example, if the player stands in the center of a large room, the player can move in any direction along the floor. (Jumping, in this room, is trivial without something to jump over or jump onto.) If, instead, the player enters a narrow hallway, there is only one meaningful axis of movement, with one direction. Even though the large room allows for more movement than the hallway, the player will move through the space in the same way: a straight path to the exit.

To create a more interesting movement space, we need more complicated level geometry, which usually means we need to pay attention to the third dimension and create meaningful elevation differences. Staircases can behave like hallways—they have a start and an end, which creates a sense of direction in the level—and they can be part of a room, giving it direction and breaking the spaces into parts. That is, when players enter a room and see distinct floors, they will begin to plan how to approach the level.

It is also possible to vary the subtlety of a level’s sense of direction. For example, in a linear level where the player’s goal is to reach a far end, you could require the player to drop down from a ledge to block the player from backtracking.

The sense of direction in multiplayer first person shooters can be more subtle, and can offer more gameplay variety. Dead ends, or one directional paths are dangerous because they reduce a player’s options to escape. If there is an item or objective to encourage the player to take this risk, these designs work like traps. A subtler form of this directional design creates variation between defensive and offensive areas. However, if there is no reason behind the directional design, or if has no role in the level, then it is bad design.

tut2_image1Some simple, directional level design

In the image above, imagine that there is a valuable item where the light object is (the sun icon). From the high ground, a player has many choices, and can easily jump down to grab the item and continue moving. However, a player who approaches the item from the low ground has to turn around; the high ground is too tall to jump onto from the low ground. This simple geometry makes one side vulnerable, and creates a subtle sense of direction between defensive and vulnerable positions.

When critiquing a level, designers often refer to “flow.” This is a vague term describing how it feels to move through the layout, and what designers consider good depends on the intended pace of the level. If you run around your level, and repeatedly pass through one room, but never enter another, then the level may have bad flow. Or, if your player becomes lost, and can’t find their way to the next area, or if the whole level is visible from the start and there is no reason to explore, then these problems also indicate bad flow. It is possible to solve bad flow with the environment art, indicating where the player needs to go, and it is also possible to place items to draw the player through the level; however, these solutions treat the effects of bad flow instead of the causes in the level itself. In levels intended for a slower pace, flow may be less important than other aspects of level design.


While constraining the movement possibilities, level geometry also occludes sightlines. A maze is a heavily occluded space where players can only see their immediate location, which informs their decisions about where to go and what to do. A large, open room is almost as bad because players can see too much at once. If there is too much to see, then this can overwhelm, or if there is nothing to see, then this can be boring. Occluding the space requires the player to move and look around in order to see everything. An interesting level should find a balance between these extremes that fits with the style of the game, and evokes the feelings you want.

tut2_image2The way level geometry opens and closes sightlines to create interesting game spaces

One important way of occluding spaces is through the use of doorways and other thresholds. These frame the next space and—as the player crosses the threshold—slowly introduces it. This framing technique also focuses on the most important part of the next area, which gives the player a sense of priority to plan around once inside the room. The plan is a kind of gestalt, where the player understands their local situation within the bigger layout.

Instead of a gradual introduction, imagine turning a corner and immediately seeing the whole space; there is no visual order or priority with a sudden change. Or imagine entering the new space from an angle where you cannot see where you are relative to a larger structure. When players don’t have a sense of the larger level, they have to advance blindly.

A space that is easy to learn and interesting to explore has some global organization or pattern, with local variation or noise. There should be a variety of open and closed, easy and difficult spaces. This kind of variation also creates the pace of a level.

In multiplayer first person shooters, the best practice for occluding the space depends on the weapon set and the player movement speed. If all of the weapons are projectile based, then players are safe when they are far from their enemies, even when they can see each other. If the weapons are hitscan—anything in the crosshair will take damage when the player shoots—then it is more important for players to keep obstacles between them and their attackers, and avoid open spaces.

Item Ecology

Items and player starts are another way to create the level’s gameplay. Items, in particular, can attract or repel the player, making some areas safe and other areas dangerous. In this way, items change the feeling of a space. It is good to pay attention to how this feeling interacts with the feeling created by the level geometry. Placing helpful items in a room that otherwise feels dangerous may create interesting encounters or traps, or it may feel contradictory. However, dangerous items in a dangerous room may feel too difficult, and may discourage a player from entering. The balance of these items should create a kind of ecology, where players have interesting options. Too many items can devalue them, and too few items can change the gameplay and the difficulty; the goal is to find the balance where items are part of a dynamic system that brings the level geometry to life.

In stealth games, items include light sources, patrolling guards, loot, and even the materials on the floor. In combination with the level geometry, these items make some areas safe and others dangerous. Too many lights and guards make stealth impossible, and too few make the level uninteresting. The balance of items in a space is what creates interesting gameplay, and problems for players to solve. In the larger scope, the variation of a stealth level between safe and dangerous areas is responsible for pacing and suspense.

In multiplayer first person shooters, the outcome of an encounter between two players depends on which is more powerful or has the advantage from a position or weapon. As the encounter develops, the weaker player will retreat to recover health, or to seek better weapons or a better position. The item placement should encourage this kind of push and pull, and create opportunities for losing players to come back. This generally means that weapons should start in the rooms where they are useful, and health items should be available to fleeing players. This also means that player spawnpoints should offer some safety.

The Limits of “Best Practices”

The goal of level design is to create opportunities for interesting gameplay. Sometimes balance is less interesting than imbalance, and sometimes the best practices lead to stale level design. Balance should mean interesting variety, and that feeling of balance emerges from the way players can move through the space, what they can see as they move, and how items create dynamic gameplay within the level.

In short, you should consider all of the advice above within the context of the game you are making. Best practices vary between genres of game, and between styles of level design. It is good to know the best practices, but it is better to develop an instinct for good design, and determine for yourself what is best.


Hopefully this article has helped you on your way to designing better levels. Thank you for reading!

Related Reading

What Happened Here? Environmental Storytelling by Matthias Worch and Harvey Smith (GDC 2010) – Slides, Recording

Meaningful Choice in Game Level Design by Matthias Worch (GDC 2014) – Video


Where Next? – Reference Manual