A moving hazard must commit its path before it enters the player's reaction window

Fact

A hazard that keeps choosing where to be while it closes on the player is not difficult, it is unavoidable. Once it is within reaction distance, any further movement of its own arrives with less warning than a human can answer, so the outcome is decided by where it happened to drift rather than by play.

The fix is not slower movement — halving the speed halves the warning needed and leaves the same hole. It is to split the approach: the hazard may move freely while it is far away, then lock onto a final track and hold it for the whole of the player's reaction window.

Why it matters

These deaths are rare enough to look like ordinary difficulty and are almost invisible in aggregate statistics — a few percent of runs, scattered. They show up as a flawless controller dying occasionally with no explanation, which reads as noise rather than as a rule violation.

They are also the deaths players remember, because the correct response was made and punished anyway.

How to apply

• Give the hazard two phases with an explicit handover distance: free movement beyond it, committed within it. Ease into the committed track rather than snapping, or the commit itself is a teleport.
• Size the handover from the worst case — maximum approach speed times the reaction budget you intend — not from how it looks at the speed you happen to be testing.
• Assert it: over a long automated run with a flawless controller, no death may occur where the hazard moved laterally inside the window. A perfect player dying at all is the signal to go looking.