Level Design For 2D Platformer Games
Given just how long 2d platformer games have been around it is perhaps surprising that no random 2D platformer level generator has ever been truly viable. However, there is a degree of balance and precision to successful platformer level design that cannot duplicated by any method other than careful design, calibration and testing.
The basics of platform level design may seem rather straightforward, but after judging a few casual games competitions and examining some platformers made in the indie designer scene, I’m seeing enough common flaws to suggest that some of the basics and principles of 2d platformer design are worth discussing.
HIGH LEVEL PLANNING
Plan your physics first
The physics of platformer games do not conform to real-world physics. The game physics by comparison to real-world physics are either stretched, or defied. The properties of avatar and weapon behavior within the game world will drive the needs of the level topography and so must be defined first. We do not want the levels to dictate the physics and gameplay anymore than a filmmaker wants a set to dictate the script.
Considerations of stretched physics include; gravity, bounciness, slipperiness, friction and buoyancy, as they relate to the behavior of the avatar and other game entities. Considerations of defied physics include; changing direction in mid-air (a la Mario), double-jumps, flying and walking on water. If missile weaponry of some sort will be included in the game, a plan for how the stretched or defied physics will effect ballistics needs to be considered.
All these considerations can be tuned and finessed later, but we should start with an initial plan.
Define your vertical and horizontal movement
For the basic horizontal locomotion of the avatar considerations include: walking, running, jumping, tumbling and skidding. For basic vertical locomotion of the avatar considerations include jumping, climbing, sliding and falling. Avoid including all possibilities for their own sake and include only what the game needs.
Define your key level components
Our key level components include the basic building blocks of the level, and how they affect physics and movement.
Define your common surface blocks; stone, metal, earth or some combination? Do any of these inhibit or enhance running or skidding?
Define your surface modifiers; ice, snow or mud? Do they add slipperiness? Affect running?
Define your dangerous surfaces; spikes or lava? Do they hurt? End the session? What about trap doors?
Define your vertical movement modifiers. Upwards movement objects may include ladders, vines, ropes, stairs, elevators or springs. Downwards movement objects may include ladders, vines, poles, slides, stairs, ropes or elevators.
As a general rule, the more heightened your physics, the fewer movement components you need. Getting up or down to hard-to-reach areas is usually a matter of either using a combination of physics and hand/eye, or accessing a movement modifier such as a ladder that has obstacles or puzzles in the way. Games usually don’t need both.
Define your game entities
Will there be enemies? How big will they be? Will any fly? Jump? Will there be weapons and tools? Power-ups? You don’t need these entities defined in detail before you start building levels (though it helps) but at the very least you need a sense of what the levels will contain, so you can build the topography accordingly.
Define your camera and filmic behavior
I recommend picking a basic navigation direction and sticking to it. Most games travel from left to right, a rule worth following since it’s familiar and comfortable to players, and moving from right to left will add nothing of value. Changing the direction part-way through the game for some levels will feel unfamiliar, again without adding anything. Even with a basically vertical level, I would still start at a left corner and end at a right corner. Many designers adopt this structure without even thinking about it. It may be worth noting that movies and TV shows almost always have their protagonists enter the frame from the left and exit to the right as well. This is part of the established pop culture vocabulary, established long-ago in live theatre direction.
Consider the kind of framing you want. How close in to the avatar is the camera? How much space around the avatar can be viewed? Is the avatar to be centered on-screen, or do you want some lead space in front of the avatar? This is an important consideration if you allow moving backwards. If the avatar can move backwards, can it navigate back to the start of the level, or just to the left edge of the screen? Can you see the topography above or below the avatar; clearly, is it hinted at, or not at all? Mock up a sample shot of what your game will look like, so you can build your topography accordingly.
PROCEDURAL PLATFORMER LEVEL DESIGN
Start with a map
It doesn’t need to be detailed, and it may even be in your head, but the best levels come from a blueprint. I sketch mine out on paper, old school, or roughly assemble a map in photoshop.
Consider the basic shape of the map. Vertical, horizontal, diagonal or square? Is it linear, or does it branch out and return to a main path? If there is a main path, design that first, and then work out any alternate, hidden routes and any dead ends off of that path. When you start building the actual level, again start with the main path and don’t fuss details off that path until the main path is done.
No wasted space
Good platformer levels have no wasted space. And no pointless dead ends. If a path leads off the main route, have something of value there; powerups, collectibles, information or a fun experience. Hedge mazes are not fun to play.
Good platformers do not make players jump to a place that does not progress the game. Avoid building topography that has no purpose. Don’t build topography and then say to yourself, “Oh, I could put some coins there.” Decide how many coins you want in that area first, and then build just enough interesting spaces to place them.
Build rhythmically (aka jumping puzzle zen)
Jumping puzzles get a bad rap. It is worth emphasizing that jumping puzzles are not inherently fun, but they can be fun when well-designed. Unless jumping is pretty much your whole game, go into each level with a plan of how much jumping there will be. Personally, I believe the best use of jumping puzzles in a level is to have either only a very few or to have a lot. In other words, to include a few jumping puzzles to break up a level that is not about jumping, or to make the level primarily about jumping and design to that.
Jumping puzzles need to feel good. And jumping puzzles feel best when the player is able to get into a sort of rhythm. Plan your complex multi-jumping puzzles like you would plan moves in a dance game, using meter. Some examples:
This does two things. One, it prevents jumping puzzles from being mere hand-eye exercises, and rewards a player for being observant, and learning as they go (as I discuss in my essay on game learning HERE, challenging hand-eye situations merely feel difficult, and hence punitive, but puzzles that are learnable give the player a reason to try again – if there is nothing to learn it isn’t a puzzle). Two, this lets players get “into the zone” to use a cliché, that when tapped into is a rewarding play experience. If you can get a player into a rhythm you can ramp up to very difficult puzzles by drawing the player up through the experience.
Lastly, jumping puzzles work best with a sense of purpose. They should take you to something worthwhile, and if they go off the main path they have to offer something of value. The best way to handle side paths with jumping puzzles is to have them loop back to the main path and not force the player to backtrack. No one likes navigating a jumping puzzle twice, and the best jumping puzzles are designed to be approached from a particular direction.
When the player launches into level 1, they should be able to understand the basic physics with regards to walking, running jumping and falling within a few seconds, so they can adjust their play accordingly. When the player launches into any subsequent levels that add or change physics the player needs to get a sense of the new physics right away. Betraying this essentially changes the rules partway through the gameplay and that leads to unhappy players who feel betrayed.
Meaningful path decisions
If you are going to present a player with a decision, make it meaningful. That is, different choices should lead to different outcomes, advantages or disadvantages. Good levels don’t use, “What’s behind door #1?” scenarios, but instead offer clues as to what decisions may offer.
One path goes up, the other down. One is dark, the other light. One is green, one is blue. Not that any of these factors should necessarily give away the results of making a choice, but players always prefer to have factors to consider, even if they really give nothing away. Choosing “up a ladder” or “down creaky stairs” is a more fun, visceral and emotional selection process than choosing door #1 or door #2. It also helps the player remember which path they took the last time when they play through again.
A path decision should not lead to a random result. For an example; a set of stairs going down leads to monsters while a set of stairs going up leads to gold, and the unfortunate player chooses down. If this player then chooses up the next time they play through, this should not lead to monsters. Taking a different path when replaying must lead to a different result, otherwise a player can end up being “randomly wrong,” twice in a row. With such a design, mastering the level would not learnable, it would be a roll of the dice.
The best-designed path decision points lead not only to different immediate scenarios, but factor somehow in the conclusion of the level, be it through the access of a weapon, power-up, sniping spot, or some other power or information. In other words, have reasons to put branching paths in your game beyond branches for their own sake. Ensure that each big decision offers a different way to beat or master the level.
Structure & Variety
Strong platformers break up the experience by offering some levels that play very differently. An underwater level, a countdown level, a flying level, and so on. You want the player to develop their skills and comfort with the game physics and entities within the first few levels, so avoid breaking up the style of play too early. On the other hand, five levels of similar play in a row can be a bit fatiguing. A good rule of thumb is to offer a varied play experience after every three or four levels.
Given how much players love discovering hidden areas, it is almost negligent to not to include some. They are inherently rewarding to discover, and they also offer replayability by offering alternate power-ups, strategies, paths, and level-completion options. Just make sure that they hold rewards of real value, and that some (if not all) also involve real challenges. If there are puzzles that you are not comfortable having in your main game (a bit too hard, or off-theme), these can be an excellent place to put them.
Put your first secret area in one of the early levels, and ensure that it not too hard to find, so players learn to look for them.
When it comes to boss battles, many platformers essentially abandon flatformer design principals and build reductionist boss battles that are pure combat games or evasion puzzles. This may date back to Super Mario Bros, the great 2d platformer template, but the boss battles were not why that game was great. The best boss levels play to the strengths of the platform game environment, and so involve motion and navigation through a space rather than reducing the game to a small room. Put bosses within areas that are no smaller than the game’s typical levels, and allow the learnings and skills the player has built to that point to apply to beating the level.