Week 6: How Forgiving Should a Jump Be?

Release cut, coyote time and jump buffer don't change the physics of a clean jump at all — they only change what happens when your timing is off.

29 March 2027

Question

If you press jump a fraction of a second before your feet actually reach the ground, should the game count it as a jump anyway? If you release the jump button a fraction of a second after leaving a ledge, should that cut your height the same as releasing it on solid ground? Neither answer is “obviously yes” — this week is about what changes when you pick one.

Prediction (skip if you’d rather just look)

Before running anything: do you expect a short coyote time window to change the maximum height a jump can reach, or only whether a jump happens at all?

Experiment

Open the Jump Lab and run an immediate-mode trial releasing jump almost immediately after pressing it, then another holding it the full charge. Compare their heights — the gap between them is what release cut factor controls: how much a short release shrinks your height relative to a full hold. Now, separately, try leaving a platform edge and pressing jump just after your character has walked off it. Whether that counts as a jump at all is coyote time; whether pressing jump slightly before landing still fires the moment you touch down is jump buffer. Both are pass/fail, not height-changing.

Explanation

Release cut factor is a continuous forgiveness: it scales how much tapping instead of holding costs you, and it’s part of the same vertical physics that gravity and launch speed shape. Coyote time and jump buffer are different in kind — they’re small windows of timing forgiveness around input and ground contact, and they don’t touch height or airtime for a jump that already happened; they only change whether a jump happens in the first place when your input was a few frames early or late. Reporting all three the same way you reported height last week would hide that difference. That’s why forgiveness numbers get reported in milliseconds and as an attempt outcome, not as a distance.

Why forgiveness windows can't be read off the trial log the way height can

The Jump Lab’s log records what a completed trial measured, not why an attempted press succeeded or failed. To measure coyote time or jump buffer honestly, you have to deliberately construct the edge case — press a beat early, press a beat late — and report what happened, the same discipline week 3 asked of any other measurement.

Task

Using the Gravity Lab, find your own working definition of “short release” versus “full hold” and report the height difference between them at your current release cut factor. Then run one coyote-time and one jump-buffer attempt, each deliberately mistimed by a small amount, and report in milliseconds how early or late you pressed and whether the jump fired.

Teaching team