Revision — memory that fades like yours
FSRS-style retention, so chapters return before they're gone.
Every measured chapter carries a retention estimate on a power-law forgetting curve — the shape large-scale spaced-repetition benchmarks find actually fits human recall:
R(t, S) = (1 + 19/81 · t/S)^(−0.5) · R(S) = 0.9
Stability S scales with how well you know the chapter — stronger memories decay slower. When a chapter's retrievability falls below 0.85 it surfaces in Revise with its freshness percentage, and the daily quiz preferentially rotates fading chapters back in. Nothing nags on a schedule; things return when your memory of them is actually at risk.
See it live — free