Study & writing
Memory retention calculator
Enter how long ago you learned something and how many times you have reviewed it to estimate how much you still remember and when to review next.
This is a rough model, not a measurement of your actual memory. Each review you log makes the memory decay more slowly.
How the estimate works
The model is based on the forgetting curve described by Hermann Ebbinghaus in 1885, in which retention falls exponentially over time: retention = e^(−t ÷ S), where t is the time since learning and S is the memory's strength, or stability, measured in days. A fresh, unreviewed memory has a low stability, so it fades fast. Each review strengthens it: here stability starts at 2 days and roughly doubles with every review, which mirrors how spaced repetition stretches the gap between successful recalls.
Worked example: you learned something 2 days ago and have reviewed it once. One review gives a stability of 2 × 2 = 4 days, so retention = e^(−2 ÷ 4) = e^(−0.5) ≈ 61%. The point where retention crosses 50% is S × ln(2) = 4 × 0.693 ≈ 2.8 days after learning, so this memory is just about to need another review to stop it slipping further.
| Days after learning (1 review) | Estimated retention |
|---|---|
| 1 day | 78% |
| 2 days | 61% |
| 4 days | 37% |
| 7 days | 17% |
| 14 days | 3% |
What the number means for revision
The lesson of the forgetting curve is timing. Reviewing just before a memory drops below about half is far more efficient than rereading something you still know well or trying to rescue something you have already forgotten. That is why the review date matters more than the percentage: catch the memory on the way down and each review costs little while buying a lot more stability. Leave it too long and you are relearning from scratch.
Treat the figures as a guide, not a readout of your actual brain. Real retention depends on how deeply you understood the material, how distinctive and meaningful it is, sleep, and how you reviewed — active recall and self-testing build far more durable memories than passive rereading, which the simple curve cannot capture. The model also assumes each review was successful; a review where you could not recall the answer does not strengthen the memory the same way. Used sensibly, it is a good prompt for when to pick something up again, which is exactly what spaced-repetition systems automate.
Frequently asked questions
What is the Ebbinghaus forgetting curve?
It is a model from 1885 showing that memory of new material fades exponentially over time unless it is reviewed. Each successful review flattens the curve, so you forget more slowly.
How much do we forget without reviewing?
A fresh, unreviewed memory can drop to around half within a day or two and to a small fraction within a week. Reviewing at the right moment resets and slows that decline.
When should I review to remember something?
Review just before retention falls below about 50%. This calculator estimates that date from the time since learning and your reviews, so you catch the memory on the way down.