How Checkpoints Work in Platform Games

A checkpoint looks like mercy. It is better understood as a difficulty control - the single knob that decides how demanding a section is allowed to be without becoming an ordeal.

The same section, two different games

Take a sequence of four precise jumps in a row. With a checkpoint immediately before it, failing costs five seconds and you try again carrying what you just learned. It is a puzzle you are working on, and working on a puzzle is enjoyable.

Move that checkpoint two minutes back and the identical sequence becomes something else. Each attempt now costs two minutes of replaying content you already solved, the learning is buried under repetition, and the frustration comes from the replay rather than the challenge.

The difficulty did not change. Only the price did. That is the entire mechanism.

Where checkpoints belong

Immediately before anything you expect players to fail more than once. Not after it, which is surprisingly common and turns a good section into a grind. Not in the middle of it, which makes the second half trivially easy on retries.

Both adventure games on this site follow the same rule: every genuinely difficult sequence has a checkpoint in front of it, and the detour to touch one is short enough that skipping it is never worth the seconds saved.

Why this frees the designer

A designer who cannot bring themselves to place checkpoints ends up softening everything else to compensate, and the level flattens into something uniformly mild. Cheap failure is what allows a section to be genuinely demanding.

This is also why neither adventure game here has a life limit. Once failure costs five seconds, limiting the number of attempts adds frustration without adding any challenge at all.

What a checkpoint should restore

Position, obviously. What is less obvious is what should not be restored, and games differ.

• Collectibles already taken usually stay taken. Otherwise a player can farm the same items by dying repeatedly, which is a degenerate strategy.
• Hazard states normally reset. A patrolling enemy returning to its starting position keeps the section predictable across attempts, which is what makes it learnable.
• Progress markers persist. A checkpoint you have touched stays touched, so you never lose ground you have already gained.

Making checkpoints visible

A checkpoint the player does not notice is worthless, and one they cannot tell they have activated is worse - they will assume they missed it and play cautiously for no reason.

The convention is a clear before-and-after state change: an object that visibly alters when touched, plus a sound. Jungle Kitty's carved stones change colour and play a distinct tone, which is enough to be unmistakable without needing any text.

Checkpoints in endless games

They mostly do not apply. An endless runner has no fixed layout to return to, and restarting is already the natural response to failure. Kitty Jump has no checkpoints for exactly this reason - a fall ends the climb, and the next climb is a new tower.

The endless equivalent is the three-second retry loop, which serves the same purpose by different means: it makes failure cheap enough that experimentation is free.

Key Takeaways

• A checkpoint sets the price of failure, which determines how hard a section can be without becoming an ordeal.
• Place them immediately before difficult sections, never after and never in the middle.
• Cheap failure is what allows a designer to make a section genuinely demanding.
• Collectibles should stay taken across a respawn; hazard positions should reset.
• Endless games substitute a three-second retry loop for checkpoints entirely.

Frequently Asked Questions

Do checkpoints make a game too easy?

No. They change what a mistake costs, not how hard the challenge is. Designers generally use that freedom to make sections harder, not easier.

Should I ever skip a checkpoint?

No. The detour is always short by design, and skipping one to save a few seconds is the most reliable way to lose a minute.

Do checkpoints save between visits?

Within a level they persist for that attempt. Completed levels and best collectible counts are saved in your browser between visits.

Why do enemies reset when I respawn?

So the section behaves identically on every attempt, which is what makes it learnable. Randomised state would make practice impossible.

Why do endless games not have checkpoints?

There is no fixed layout to return to, and restarting already costs only a few seconds.