📖 Overview

Learning fades even after strong first comprehension.

This calculator projects remaining retention from initial mastery and daily decay assumptions.

Use it to schedule refresh sessions before forgetting crosses critical threshold.

🧪 Example Scenarios

Use these default and higher-pressure example inputs to explore how sensitive this calculator is before using your real numbers.

InputBase CaseHigher Pressure Case
Initial Retention (%)92110.4
Daily Decay (%)2.32.76
Days Since Last Revision1012

⚙️ How It Works

Estimates memory decline over time with a daily exponential decay model for review planning.

The Formula

Current Retention = Initial Retention × (1 − Daily Decay Rate)^Days
Initial RetentionEstimated mastery immediately after learning/review
Daily DecayAverage daily forgetting rate assumption
DaysTime since last substantial revision session
Current RetentionEstimated retained knowledge now
💡This model helps timing decisions for spaced repetition; it is a planning approximation, not a neurological measurement.

Quick Reference

InitialDecay/day10 days20 days
90%1%81.4%73.6%
90%2%73.5%60.1%
95%2.5%73.3%56.6%
85%3%62.7%46.2%

When To Use This

  • Use this tool when you need a fast decision during active planning or execution.
  • Use this before committing money, time, or tradeoffs that are hard to reverse.
  • Use this to compare options using the same assumptions across scenarios.

Edge Cases To Watch

  • Results can be misleading if key inputs are missing, stale, or unrealistic.
  • Very small or very large values may amplify rounding effects and interpretation risk.
  • If assumptions change mid-decision, recalculate before acting.

Practical Tips

💡 Use this to trigger spaced review before retention drops too far.
💡 Higher initial retention does not eliminate decay pressure.
💡 Tune decay rate from your own recall-test history for better accuracy.

Frequently Asked Questions

❓ Is daily decay constant in real life?

No, but a constant-rate model is useful for planning.

❓ Can retention go below zero?

No, the model asymptotically approaches zero.