WordieForgetting Curve Calculator
Enter how long ago you learned something and how many times you have reviewed it. See your estimated retention, a forgetting curve with and without reviews, and how much survives after a week, a month and three months.
What does the forgetting curve calculator show?
This calculator estimates the chance that you can still recall something, such as a new word, a given number of days after you learned or last reviewed it. It uses a simplified Ebbinghaus-style model: memory fades quickly at first and then more slowly, and every successful review makes the next decline slower. The two curves show the same item learned once and never revisited, and the same item after the number of reviews you entered.
Hermann Ebbinghaus measured his own forgetting of nonsense syllables in 1885 and found the steep early drop that gave the curve its name. Later research on the spacing effect and retrieval practice explains why reviews flatten it. For the full story, read the forgetting curve explained.
The forgetting curve formula used here
The calculator uses the exponential form that most modern descriptions start from:
R = e^(−t / S)
Ris retention, the estimated probability of recall (0 to 1, shown as a percentage).tis the number of days since you learned the item or since your last review.Sis stability, measured in days. Whent = S, retention has fallen to about 37%.
Each review multiplies stability by a gain factor g, so after n reviews S = S₀ × gⁿ. With the defaults (S₀ = 2 days, g = 2.5, three reviews) stability grows from 2 days to about 31 days, and retention one week later is about 80% instead of 3%. The gain of 2 to 2.5 per review is an assumption chosen to mirror the way spaced repetition intervals typically grow; it is not a measured constant.
How to interpret your retention estimate
| Estimated retention | What it means | What to do |
|---|---|---|
| 90% or more | Fresh, easy to recall | Reviewing now adds little. Wait longer. |
| 70–90% | The useful review zone | Review soon. Recall takes effort, which strengthens memory most. |
| 40–70% | Fading | Review today; expect some misses. |
| Below 40% | Mostly forgotten | Relearn it, ideally with a fresh example sentence. |
The stat "until recall drops to 80%" is a practical guide to your next review date. Many spaced repetition systems aim to schedule each review when predicted recall falls to somewhere around 80 to 90%, so the effort of remembering is real but success is still likely. The spaced repetition scheduler turns that idea into calendar dates.
Limitations of this model
- It is a simplification. Researchers have fitted power functions, sums of exponentials and other shapes to forgetting data. A single exponential is easy to understand, but it is not the only defensible choice.
- The numbers are assumptions. Real stability depends on the word, your prior knowledge, sleep, interest and how deeply you processed it. A word tied to a vivid sentence lasts far longer than one read in a list.
- Reviews are not all equal. A review where you actively recall the answer strengthens memory more than rereading it. A review done too early, while recall is still near 100%, adds less than the model assumes.
- Failed reviews are ignored. The calculator assumes every review succeeded. Modern schedulers such as FSRS model failures, difficulty and stability separately.
Related reading
- Forgetting curve: definition
- A practical guide to spaced repetition
- How to stop forgetting new words
- What the spacing effect research says
How to use this tool
- Enter the number of days since you learned the word or since your last review.
- Enter how many spaced reviews you have already completed.
- Optionally adjust the initial stability (how many days a brand-new memory lasts) and how much each review strengthens it.
- Read your estimated retention, then compare the two curves and the 1, 7, 30 and 90 day table.
Frequently asked questions
What is the formula for the forgetting curve?
A common simplified form is R = e^(−t/S), where R is the probability of recall, t is the time since learning and S is memory stability in days. Larger S means slower forgetting.
How much do we forget after one day?
In Ebbinghaus's own experiments with nonsense syllables, a large share was lost within the first day. For meaningful words the drop is usually smaller, and it depends heavily on how well the item was learned in the first place.
How many reviews does it take to stop forgetting a word?
There is no fixed number, but each successful review makes forgetting slower. In this model, three or four well-timed reviews stretch a memory from days to months.
Is this calculator scientifically exact?
No. It uses a simplified Ebbinghaus-style exponential model with assumed numbers. Real forgetting varies with the person, the material and how it was studied, so treat the output as an illustration, not a measurement.
Does this tool store my data?
No. All calculations run in your browser and nothing is uploaded or saved.
Turn these numbers into real words
Wordie quietly builds the habit for you: save words while you read, see them highlighted everywhere, review in seconds.
Add Wordie to Chrome — it's free
