This essay is meant to address the game design equivalent of pacing in films. In films there is establishment, development, twists, and conclusions. The 3 act structure. In games there is a progression between the point where you don’t understand anything of a concept, where you know a bit of the concept, and where you are so familiar with a concept that you can have your predictions fooled or distorted about a new situation.
Phase 1: Teach
Phase 2: Challenge
Phase 3: Subvert
Level design is the essential process of progression through a game. Its intent is to lay bear how the game works, test the player’s mastery of it, then try to destroy the player by all means possible once the player understands how to combat it. Good level design can be judged by the phase of the game, and by how well it constrains the player while also letting the player play off it. Phase 3 is the type of level design that we want to aspire to, but the other stages must exist because a player must learn. They can’t play at the top level immediately. This is why difficulty curves exist, to bring the player up to the top. From this point of view, level design is a process of crushing the player via any means possible, and making them accept their defeat. It’s an iterative process of building the player up so you can fairly send them to fall. To use the classic metaphor, give them enough rope to hang themselves with.
Portal and especially portal 2 is a game that never moved past phase 1, and their dev commentary expresses this. Portal 1 later on got into phase 2, and the challenge chambers satisfied phase 3. Portal 2 lacked at anything past phase 1
To a smaller extent, every area and new mechanic has to go through these phases, where first the player is taught about the new thing, the lay of the land, the new enemies, the new tools, and how they’re used or work, then they go through challenging the player, and
Beyond that, level design, and game design in general is a process of both giving the player things to work with, and making it hard on him. In general, the thought goes, how can you make this harder on the player in an intellectual way? The easiest way to make something hard is just to give the player almost no ammo and tons of enemies in tiny space without options for escape, but this isn’t interesting, it isn’t smart for the player or developer, and it’s not good level design.
A better thought process is to think about giving the player options, and the enemies means of countering these options. The difficulty comes from using the environment to the advantage of the player, from making the correct choices given the tools offered. To both succeed in terms of manual dexterity, and in terms of intellectual problem solving. Games that aim for one or the other often get neither. Beyond this, a good way to think about level design is as a process of giving the player things to do, ways to succeed, and then thwarting them, so that the player must dig deeper to try to find viable strategies. If there is a common way to exploit enemies, instead of just removing it, leave it in but give enemies a way to fight back. If there is a spot on the map where the player can stand and never get hit, then either trim it out, or put a firepit there, or make it so the player can’t hit anything while there either. If the players can easily avoid enemies by just ignoring them and running past them, then make the enemies chase them or block off exits. If you give the player the option to heal then have enemies try to close in if the player tries to get in a safe spot to heal.
A great recent example of this sort of thinking would be hotline miami which is heavily based on taking out enemies and provoking them. Enemies carry guns and melee weapons. You start out each level with just your fists. Your fists are silent, and you provoke enemies by being seen. Your fists however are inefficient, and you must pummel the enemy on the ground to guarantee they will not attack you again. Melee weapons generally have a better range than fists and kill instantly, and can be thrown to stun enemies. They are also generally silent. Guns have limited ammo, and are generally trickier to aim than melee weapons, but they kill and can do so at a distance, but the cost is that they make a loud sound which can provoke enemies that cannot see you.
Notes for expansion:
twists early on disrupt learning and consistency in the learning environment
owata style games are universally simple, so that they can throw tons of twists
Alternate proposal for phases: