Independent games frequently do one thing and do it for several hours. That is a structural decision rather than a limitation, and it changes how the whole game is designed.
Scope forces a choice about breadth or depth
A large studio can staff several systems in parallel, so a game can have combat, crafting, dialogue and traversal each built by a different group.
A small team cannot. Every additional system competes for the same people, and a game with four half-finished systems is worse than one with a single finished one.
So the design question becomes which single interaction is rich enough to sustain the whole running time, and everything else is arranged to serve it.
One verb generates a surprising number of situations
A mechanic is not a single behaviour. It is a rule, and a rule interacts with whatever the level places around it.
A mechanic that moves the player also moves them into hazards, past enemies, through timed gaps and away from things they need, and each of those is a different puzzle.
The designer's work is finding those interactions rather than inventing new mechanics, which is a far cheaper form of content to produce.
Levels become variations on one question
Once the mechanic is fixed, a level is an arrangement that asks the player to understand a new consequence of it.
Good progressions introduce a consequence in isolation, then combine it with something already learned, then remove a crutch the player had been relying on.
This is why these games feel like they are teaching without tutorials. The teaching is the level order, and the player is never told what they have just learned.
Difficulty comes from combination rather than addition
A game with many systems raises difficulty by stacking numbers or adding enemies, because the systems themselves are already fully introduced.
A single-mechanic game raises it by requiring two known consequences at the same time, under pressure, which demands understanding rather than reaction.
The result is a difficulty curve that feels earned. Failure points at something the player has not yet grasped, which is a more useful signal than a health bar that is simply longer.
The approach has a natural end point
Every mechanic eventually exhausts its interesting combinations, and pushing past that point produces levels that are merely fiddly rather than clever.
Experienced designers plan the running time around that limit, which is why these games are often short and rarely apologise for it.
Stretching them is what breaks them. Padding a single-idea game with filler undermines the one quality that made the idea worth building a game around.