Learning Retention Calculator
Compute learning retention rate as (retest score / initial test score) × 100, telling you what percentage of original mastery has survived the gap. The simplest forgetting-curve metric; useful for evaluating spaced-repetition schedules and study-method effectiveness.
Last updated: May 2026
Compare with similar
About this calculator
Retention rate = (retest score / initial test score) × 100. A retention rate of 100% means perfect retention (you scored the same on the retest as the original test); below 100% indicates forgetting; above 100% indicates learning between tests (rare without continued study). Variables: initialKnowledge is your initial test score as a percentage; retestScore is your retest score on the same or equivalent content after some elapsed interval. Edge cases: initialKnowledge must be > 0; both should be expressed in the same scale (typically 0-100 percent). The formula doesn't capture how much time has passed — a 95% retention after one week is very different from 95% after one year. Hermann Ebbinghaus' classic 1885 forgetting-curve research found that without active review, memory of nonsense syllables decayed exponentially: ~58% retention after 20 minutes, ~44% after 1 hour, ~33% after 9 hours, ~28% after 1 day, ~25% after 2 days, ~21% after 31 days. Meaningful material decays much more slowly than nonsense syllables, but the same exponential shape holds. Spaced repetition systems exploit this by scheduling reviews just before forgetting is predicted to occur, dramatically improving long-term retention with minimal review time. Modern algorithms (FSRS, SM-17) personalise the curve to each card and each user. The retention rate metric is most useful as a rough check: if your retention after 1 month is below 60%, you need more frequent review or better study methods; above 85% suggests intervals can be lengthened to save review time. Reference: well-spaced Anki users typically achieve 85-90% retention rates on their daily reviews, indicating intervals are appropriately tuned.
How to use
Example 1 — Cramming retention. You scored 95% on a test, then took the same test one month later without studying in between and scored 80%. Enter Initial Test Score = 95, Retest Score = 80. Retention = (80 / 95) × 100 = 84.2%. ✓ Modest forgetting over a month — common for actively-engaged learning; would be much worse for purely passive memorization without retrieval practice. Example 2 — Poor retention indicating cramming. Scored 95% on a final, but a follow-up assessment six months later shows 40%. Enter 95, 40. Retention = (40 / 95) × 100 ≈ 42%. ✓ A retention rate this low after six months suggests the original 95% was largely surface knowledge (cramming for the test), not deep understanding. Most of the material wasn't consolidated into long-term memory. Compare: students taught with spaced practice and retrieval testing typically retain 70-85% at six months on the same material. The same student could probably regain the lost knowledge in 2-3 hours of review, but the forgotten material is functionally unavailable for new learning that builds on it.
Frequently asked questions
What is the forgetting curve and why does it matter?
The forgetting curve, first measured by Hermann Ebbinghaus in 1885, describes how memory of newly learned material decays over time without review. Ebbinghaus found that retention of nonsense syllables drops dramatically in the first few hours and days: ~50% lost after 1 hour, ~70% lost after 24 hours, ~80% lost after a week. Meaningful material decays much more slowly but follows the same exponential pattern. The forgetting curve matters because it predicts when review is needed: trying to remember at the right moment (just before forgetting) strengthens the memory more than reviewing too early (wasted effort) or too late (already forgotten and must be relearned). Spaced repetition algorithms (SM-2, FSRS, Anki defaults) explicitly model the forgetting curve for each card and schedule reviews at the calculated optimal interval, typically yielding 85–90% retention with much less review time than massed practice would require.
How does spaced repetition improve retention?
Spaced repetition exploits the "spacing effect" — discovered by Ebbinghaus and confirmed by hundreds of subsequent studies — that memories formed with rest intervals between study sessions are dramatically more durable than memories formed by massed (back-to-back) practice. A flashcard reviewed once today, again tomorrow, again in 4 days, again in 10 days produces far better long-term retention than the same card reviewed 4 times in one session. Modern spaced-repetition algorithms automate this: each card has its own scheduled review date, computed from the user's past performance on that card. If you remember a card easily, the next interval grows (perhaps 1.3-2×); if you forget, the interval resets and you relearn. Over months and years, this produces "permanent" learning of even very large bodies of material (10,000+ flashcards for a medical school deck, 5,000+ for a language) with as little as 15-30 minutes of daily review. The investment is daily consistency, not session length.
What's a good retention rate for spaced repetition?
Most spaced-repetition optimisers target 85–90% retention rate on the daily review queue. Lower than 85% suggests intervals are too long (too much forgetting between reviews); higher than 92% suggests intervals are too short (wasted effort). Anki's default settings tend toward ~85%, while FSRS (a newer algorithm) aims for ~90% by default with personalised tuning. Lower targets (70–80%) mean fewer reviews but more frustration from forgotten cards; higher targets (95%+) mean very smooth reviews but excessive total review time. The right target depends on what the material is used for: medical-school decks where missing a fact can hurt patient outcomes should aim for 90%+ retention; casual vocabulary learning can target 80% with less anxiety. The 85–90% range is a sweet spot balancing efficiency and reliability for most learners.
What are the most common mistakes interpreting retention rate?
The first is ignoring the time interval — retention rate of 80% means different things at 1 week vs 1 year, but the formula doesn't encode that. Always note the elapsed time. The second is treating retention rate as the only success metric; depth of understanding, ability to apply, transfer to new problems all matter more than verbatim recall. The third is using the same test twice; students may remember the test items themselves rather than the underlying concepts, inflating retention. Use parallel forms (different questions, same concept) for valid retest comparisons. The fourth is confusing retention with re-learning — material you've "forgotten" can usually be regained much faster than initial learning, so a low retention score isn't total loss. The fifth is assuming retention is uniformly distributed across the material; some facts stick easily (high-interest, well-explained), others fade quickly (arbitrary, abstract, isolated). Aggregate retention hides this variation.
When should I not use this calculator?
Skip it for skill-based learning where the relevant metric is performance, not recall — you don't test bike-riding "retention" via a quiz; you ride a bike. Sports, music, hands-on skills, and procedural knowledge need performance assessments, not test scores. Don't use it without controlling for test-question quality; the same nominal "test" can have wildly different difficulty between administrations. It's the wrong tool for evaluating teaching effectiveness in classrooms; cohort-level retention rates need much more statistical care (matched groups, control for variation, demographic factors). Avoid it for short retest intervals (< 1 day) where the result is largely about consolidation rather than long-term retention. Finally, don't use it to compare across people without controlling for baseline ability, prior knowledge, and interest; retention is highly individual.