Level Design is Game Design
This essay is a work in progress. Last updated July 22, 2024
Level design is game design, and game design is level design.
If you’re making video games, especially 3D video games where the player is taking on the role of a single character and controlling their actions in the world – then the only way for the player to experience anything at all is to place the player in a level – and have things to do in that level. Without level design Super Mario is a cartoon plumber floating in a black void – with nothing to stand on or jump onto. Without level design Lara Croft is a plucky explorer with no ledges to grab, wolves to shoot, or tombs to raid. And without level design Niko Bellic of Grand Theft Auto 4 is just a guy with no cars to steal and nowhere to commit crimes.
In many games the player can only experience a single level at a time. In Pac-Man each level is a maze that fills the entire screen, and when the player completes one level, they proceed to the next new level that fills the screen. In Super Mario Bros. each level is multiple screens worth of environment, and once the player runs and jumps from the starting point on the far-left side of the level all the way to the far-right side of the level, the level is completed, and the player proceeds to the next level.
But linear games and linear levels are not the only way to structure things. In The Legend of Zelda, the player can explore and move through the world in any of the four cardinal directions, with each segment or “room” the player visits filling the screen. In many ways the entire game is one big level, with discreet sub-levels when the player enters caves and dungeons. And at the same time each individual room of the game is its own descreen “level” that adheres to the rules and conventions of level design.
Without level design these games simply would not exist. And if any individual power-up, trap, weapon, or enemy isn’t placed somewhere inside these levels, then they would not appear in the game at all. Levels define what the player can experience and interact with in these games, and as such the act of making levels is making the entire game. And all of the concepts and fundamentals of good game design apply directly to good level design.
This is why I say level design is game design, and game design is level design. Level design is where the rubber meets the road, and it’s up to a level designer to take all of the assets and game mechanics and features created by the rest of their game development team, and make sure that they are placed and sequenced in a fun way to be experienced by the player.
Games are a set of challenges and rewards, governed by a set of rules, and for any level of a game to be fun itself, it has to have all of these key elements of a game.
Let’s discuss how each of these elements come into play in level design, starting with the rules.
Level design usually begins with the creation of the physical environment – the objects that make up the floor, walls, and ceilings of the space. The player needs somewhere to stand and somewhere to go, after all. Since most of these elements in the level cannot be moved, they very much define the rules of the level:
“You may move here, you cannot move there, you may jump up here, you may crawl under there.” When combined with the player character’s movement abilities and the game’s simulation of gravity, the level is a set of HARD rules that the player can physically come up against.
Making some of the objects in the level move also defines rules. The designer can define a rule that a platform moves back and forth, that an elevator moves up and down, and that doors open and close. Sometimes these objects can be set to move on their own, moving along a looping path or back and forth forever. These predetermined movements of the physical objects in the level are miniature game rules specific to the level.
Scripting in the level can be used to further define rules of play. The level designer can connect different event inputs and outputs to make the level feel like a believable world. The inputs can be things like the player walking into a specific area, the player activating a switch, or the player defeating a specific enemy. The outputs of these triggers can be opening a door, activating an elevator, spawning a new tougher enemy, or any number of other actions.
But the placement of objects in a level doesn’t only define rules, it also defines many of the challenges the player will come across. The layout and placement of walls in a level define how easy it is for the player to find their way from the entrance to an exit. Is the level a wide-open space where you can easily see the exit and figure out how to get there? Or is it a cramped maze with no hints of where to go next?
The objects in the level also define how easy or challenging it is to physically traverse the space, making progress towards the exit as the player walks, runs, jumps, or drives from beginning to end. If the game is a “platformer” like Super Mario Bros. that requires the player make a series of jumps between platforms to reach the end of the level, then everything about where those platforms are placed in space define how easy or challenging the level is.
If platforms are placed close together, the challenge to jump between them is easier. If platforms are placed far apart, then the jump between them is more difficult. If a platform is large it’s easy to land on, and if a platform is small it’s more difficult to land on. And you better believe moving platforms are even more difficult to safely land on. The selection of objects, where they’re placed in relation to each other, and the sequence that the player is asked to move between them, all add up to defining the physical challenges of the level.
Furthermore, designers can place additional physical challenges in their levels that the player must avoid or overcome to proceed. Traps can be placed in the level that damage or immediately kill the player when touched. There can be pits filled with spikes, lava, or acid that the player has to jump over. And there can also be gaps or holes in the floor, revealing bottomless pits that the player can fall into… to their doom.
Level design is not only about making the game challenging for the player. Games are only fun when the player is also getting something in return for playing. Since things can only be in the game if they are actually in the levels, it’s up to level designer to make the levels themselves rewarding and therefore fun.
One way that levels can be rewarding is by looking really cool. They can contain buildings with inspiring architecture, they can feature breathtaking natural scenery like waterfalls or majestic forests, and they can have grand vistas that inspire the player to explore. Video games are a visual medium, and making a level look cool and visually distinct can sometimes be enough to make the player want to spend more time there.
But more than the visuals in a level can make it rewarding. The act of standing somewhere up high and looking down on other parts of the level can be rewarding. When you’re on the top floor of a skyscraper in real-life, it can make you feel powerful and good about yourself. Similarly, when a level gives you a calm moment to look back at the terrain you’ve had to trek across to get there, that can be deeply satisfying. You can feel good about yourself as you reminisce about the grit it took to get here, and maybe even brag to friends with a screenshot from that vantage point.
We’ve only talked about levels as a hard physical world so far. We haven’t discussed enemies or items or fog or anything else besides making a place for the player to explore. And that’s because the physical world alone can make a level fun. The space you navigate can be challenging, it can be rewarding, and it absolutely has its own set of ground-rules (no pun intended).
Games where the primary challenge is running and jumping from platform-to-platform, or “platformers” for short, are a favorite game genre of mine. It’s fascinating to see how the genre continues to evolve and grow! From the Donkey Kong to Celeste, platformers highlight how level design can most closely define the overall fun of an entire game.
Platformers are also alive and well in the modding and user-generated content scene. People are making their own platformers in Minecraft, Roblox, and Fortnite, using the tools in each of those game’s level design toolbox to select and place platforms in fun and interesting sequences that challenge players. Even the name of this genre is evolving, and creators list these levels under terms like “parkour”, “obstacle course”, “obbys”, “deathrun”, “only up”, and more!
The barrier to entry for creating platformer levels is incredibly low, anyone can do it. But at the same time, mastering the art of platformer level design takes lots of practice, playtesting, and iterative creation.
When designing platformer levels, the degree of challenge of a level can be measured by the sum of the difficulty of each individual jump and maneuver the player has to make. You can even think of each individual “jump” as a teeny-tiny game of its own. The challenge of the jump is defined by a number of level design factors, and the reward of the jump is defined by what’s on the other side of the jump.
The simplest jump in a platformer is one that only requires the player to press the jump button, and do nothing else. These jumps are common in 2D platformers where the player can walk in front of a platform, and while standing still, jump up onto it. An early example of this type of platform is in Super Mario Bros. 3. But the ability to jump directly up “through” a platform is uncommon in 3D platformers.
Meanwhile in 3D platformers the most basic jump is one where the player can walk directly up to a short wall (or curb) that they cannot automatically walk over, move in the direction of the wall, and while doing so, press the jump button. Hop! The player jumps up and over the curb, and lands safely on top of the wall. A slightly more difficult jump would be a little higher, and likely require that the player hold the jump button down a little longer.
But there is nothing at risk attempting these jumps – there’s no gap between the “launchpad” where the player starts, and the “landing-pad” where the player is trying to get. This would barely register as a challenge to most people, and seems completely trivial to us game designers. But for beginning levels or tutorials, this technique is often used to teach the player the game’s mechanics. Halo 3 opens with the player in a wooded area, and they have to do a simple jump like this to hop over a log to continue down a path and into the game.
There is an almost equally trivial jump to discuss, the flat jump between platforms with a small gap between them. Imagine you have two platforms practically next to each other on the same vertical plane, but with just enough room between them for the player to fall through between them. We now have something at risk! The player can not only fail the jump, but there is a penalty for doing so. The player falls down and at a minimum they have to navigate back to the launchpad to attempt the jump again. But the penalty could be worse – there could be spikes or lava or nothingness to contend with.
Being forced to backtrack is often enough to make the player not want to fail an attempted jump. The further the player has to walk back to try again the worse the penalty will feel. And if the player has to redo multiple jumps to catch back up to where they were when they failed, the penalty will feel that much more severe. But it is this escalation of risk that makes any individual jump feel like it has stakes. If the player has something to lose, like the time they’ve invested into the game to get this far, then they will be that much more excited when they succeed!
Jumps can be made more and more difficult by increasing the size of the gap – moving platforms further and further apart – horizontally and/or vertically. The further apart the platforms are, the more difficult the jump will be and the more skill and precision it requires from the player. There will ultimately be a horizontal, vertical, and diagonal distance between platforms that is completely impossible – because the player character just can’t jump this far.
Besides the distance between platforms, the next most important way to make individual jumps feel different is to change the size of the launchpad and landing-pad in the jump. As discussed earlier, the size of the landing-pad particularly affects how difficult a jump is. A small landing-pad requires extra precision from the player to successfully make the jump. Meanwhile, if the landing-pad is large and has guardrails preventing the player from overshooting their target, then the jump will be much easier.
Sometimes a particularly long jump requires that the player have a running start to make the jump. This can be challenging if your launchpad is small, and sometimes impossible if the launchpad is tiny. So instead the player might be required to not only have a running start, but also to be running at top speed through a series of jumps, with this tiny platform being a spot they only briefly land on before immediately making their next jump… like an Olympic triple-jump.
And when you’re working on 3D platformers, the depth and width of your platforms come into play. Jumping between platforms that are 10 feet wide can feel pretty comfortable, while jumping between six-inch-wide balance-beams feels super dangerous and thrilling. Having a variety of platform shapes and sizes will not only make your level more visually interesting, it will also give the player a variety of challenges to engage with.
While it’s interesting to discuss the relative difficulty of challenges in a level, these distinctions really become important when you move beyond designing an individual challenge or jump, and start considering the sequence of challenges in a single level. The task of designing this sequence of challenges in levels reminds me of composing music – picking the right notes and chords to play at each moment, and arranging them in a satisfying way.
We can borrow the concept of sheet music and apply it to level design, letting us look at a level as a sequence of symbols on paper instead of playing them in a game or looking at them in a level editor. Remember that only playtesting your levels with real players will tell you how actually challenging, rewarding, and fun they are. But this paper notation can be useful for discussing different styles of level design, and as a way to promote important concepts like variety and pacing in your designs.
Let’s imagine that we give each square meter of our game a single-digit challenge-score between zero and nine, with zero meaning “there is no challenge, the player is completely safe” and nine meaning “the player is stressed out and losing their mind with how difficult this is.” Then we can investigate some hypothetical level designs by looking at a sequence of numbers. And while they might “look” like a simple 2D level, most 3D platformers can still be boiled down to a sequence of challenges, just with the player having to turn every now and then between challenges.
Here’s a level that’s completely flat with no jumps or gaps at all:
00000000000000000000000000000000
This could be the worst level ever… but there’s one worse: a level of only extreme-difficulty challenges:
99999999999999999999999999999999
Not only would that level be too difficult, but it would also be boring! Variety and pacing are important parts of game design, and you want to mix it up to keep the player’s attention. How about this level?
00001112223334445556667778889999
So this level is more interesting than the first two, and it certainly has variety. But it’s unrelenting, it never gives the player a chance to catch their breath, it’s just ever-increasing difficulty from start to finish. And this consistency makes the level predictable, and that predictability can cause a different type of boredom in the player.
So let’s try something like this:
00100100200100200200300300200400
You’ll notice this level still has an escalation to it, with the player presented a series of challenges between large gaps of safety. The player can catch their breath and there’s some predictability to the pattern, but by the end we mix it up and surprise them a bit with a head-fake, presenting them with the easier 2 right before hitting them with the 4 at the end.
01010202010303020404030505040600
This is a faster-paced level with a steady amount of safety between challenges, but the challenges are more tightly packed together and ultimately get more difficult. But a level doesn’t have to follow a single pace or tempo throughout. We can make a level “speed up” by the end, making it feel even more exciting, like this:
00010002002003020303012023023400
Experiment with your own number sequences, or try documenting the sequence of one of your favorite platformer levels, and see what you can learn from the experience.
When sitting down to consider just how difficult any one jump in a level is, you can use this list of factors to give your jumps a relative score. Each game will be slightly different, but I’ve tried to include the elements found in almost all platformers, be them 2D or 3D.
A jump challenge can have any number of these factors:
It’s relatively straightforward to design the first level of a game, with a minimal amount of challenge and risk involved – and it’s easy to recruit new players who have never played the game before to try it out. And it’s also pretty easy to design really difficult levels, because the development team will be so good at the game that what they find to be challenging should only appear at the very end of the game.
But it can be a real challenge to design and balance levels in the middle of your game, and that requires a serious commitment to long-term playtesting with repeat players, who can play through the entire game, and give you feedback based on their experience of playing through lots of the other levels, but have not yet developed skills to rival the development team.
There certainly is more to a level than the physical objects that the player can walk on, jump over, and bump into. And while level design usually begins with the physical world, there are many more types of elements that we can place in our level designs to make it an even more fun and memorable experience.
We sometimes talk about games as collections of assets – lots of individual components that add up to an entire game. Each level is a particularly complicated type of asset, including or pointing to dozens (if not hundreds or thousands) of other assets.
What types of things make up a single level of a game?