Photo by Javier Martínez on Unsplash. Illustrative controller photograph, not Celeste footage.
Celeste lets you jump briefly after leaving a ledge. It also remembers a jump pressed and held shortly before landing, so the jump can happen when you touch down.
Those two rules—coyote time and jump buffering—are more than conveniences. They help shape the relationship between a player's intention and the movement they see. A screenshot cannot show that relationship. Playing can.
That is a useful starting point for a design question: what if we reviewed a movement mechanic with the same care we give an illustration?
The mechanic is part of the aesthetic
In Celeste & Forgiveness, Maddy Thorson describes these and other small allowances. Thorson explicitly does not claim to have invented them. The point is how they work together in Celeste, widening timing or positioning windows in the player's favor.
Notice the distinction between difficulty and input interpretation. You can keep a demanding route while being generous about whether a near-edge button press counts as the jump the player intended. That does not mean every game should use the same tolerance. A particular game's movement, level geometry and intended challenge determine what makes sense.
This connects to an older philosophical discussion worth revisiting. Wesley D. Cray's 2018 review of The Aesthetics of Videogames describes Zach Jurgensen's argument for attending to games as games, including the aesthetic evaluation of mechanics. It also discusses Jon Robson's attention to particular player performances.
The book is from 2018, not a new 2026 release. I am using that perspective as a design lens, not presenting a new experiment or claiming the book analyzed these specific Celeste techniques.
A small graybox comparison
Here is a proposed exercise for a platformer prototype. It is a playtest plan, not a report of tests I have run.
Build one short route with an edge jump and a landing followed by another jump. Keep the camera, speed, geometry and presentation unchanged across three variants:
- Strict: jumping requires being grounded when the fresh input arrives.
- Ledge grace: add a short post-ledge window; keep landing input strict.
- Ledge grace plus buffering: also retain a recent jump request briefly so it can be consumed on landing.
Use your engine's existing movement and collision loop. Expose the two windows as separate tuning values rather than bundling them into one “forgiveness” slider. There is no universal duration supplied here, and no claim about Celeste's exact timings.
For each attempt, record the input time, grounded-state changes, whether a jump was accepted, and the player's own account of what they tried to do. If you retain input for buffering, consume it when the jump fires; otherwise one press can accidentally trigger repeated jumps. Clear stale requests when restarting a trial.
Review intention, not just completion
A completion count tells you whether someone reached the end. It does not, by itself, tell you whether the movement felt expressive, predictable or satisfying.
Ask the player to describe a specific moment: “What were you trying to do at that edge?” Then compare their account with the input trace. A rejected press might be intentional difficulty, an unclear visual boundary, or an input window that feels at odds with the action. The trace helps locate the event; it does not read the player's mind.
Vary the order in which people try the versions. Repeating the same route teaches it, so a faster later attempt is not automatically evidence that a new setting is better. This small exercise can generate useful observations, but it cannot establish a universal preference or a cognitive benefit.
Give player expression room to appear
The payoff is not simply more successful jumps. It is a clearer conversation about the kind of movement you want to make possible: deliberate, flowing, tense, playful, or something else.
Art, sound and animation matter enormously. Mechanics also deserve aesthetic attention before those layers arrive. A graybox can already invite a distinctive kind of performance.
Which movement detail would you test first: ledge timing, landing inputs, or control near the top of a jump?
Read the KRI ZEK philosophy-of-play article.
Download Altered Brilliance: https://play.google.com/store/apps/details?id=tech.krizek.alteredbrilliance
Global website: https://global.krizek.tech
Pre-register for Arzenal Human Health: https://play.google.com/store/apps/details?id=tech.krizek.arzenal
Join The Power Of Gaming: https://discord.gg/sbYSPcCqJn
Top comments (0)