What is FSRS Desired Retention and what should I set it to?

Desired Retention is the target recall probability when a card appears for review. The official Anki manual establishes 0.90 (90%) as the default standard, while 0.85 to 0.90 is widely recommended for high-stakes medical exams like USMLE Step 1 and Step 2 CK to prevent review fatigue. During intense clinical rotations, reducing desired retention to 0.84–0.85 decreases daily review volume by approximately 30%–35% based on FSRS interval scaling formulas.

Why does increasing retention from 90% to 95% cause review overload?

Human memory follows a power-law forgetting curve where retrieval probability decays over time. Chasing the final 5% of retention requires intervals to contract dramatically. In FSRS modeling, increasing desired retention from 90% to 95% nearly doubles daily review volume for a marginal 5% gain in factual recognition.

How do I change Desired Retention in Anki Desktop?

Open Anki Desktop, click the gear icon next to your deck, click Options, scroll down to the "Advanced" section where FSRS is enabled, locate "Desired Retention", enter your target decimal (e.g. 0.88 or 0.90), and click Save.

Do I need to change learning steps when using FSRS?

Yes. Under FSRS, learning steps should be short and intra-day (e.g. "10m" or "15m"). Multi-day learning steps like "1d 3d" should be avoided because FSRS calculates optimal inter-day intervals starting from Day 1.

What is FSRS Desired Retention and what should I set it to?

Desired Retention is the target recall probability when a card appears for review. The official Anki manual establishes 0.90 (90%) as the default standard, while 0.85 to 0.90 is widely recommended for high-stakes medical exams like USMLE Step 1 and Step 2 CK to prevent review fatigue. During intense clinical rotations, reducing desired retention to 0.84–0.85 decreases daily review volume by approximately 30%–35% based on FSRS interval scaling formulas.

Why does increasing retention from 90% to 95% cause review overload?

Human memory follows a power-law forgetting curve where retrieval probability decays over time. Chasing the final 5% of retention requires intervals to contract dramatically. In FSRS modeling, increasing desired retention from 90% to 95% nearly doubles daily review volume for a marginal 5% gain in factual recognition.

How do I change Desired Retention in Anki Desktop?

Open Anki Desktop, click the gear icon next to your deck, click Options, scroll down to the "Advanced" section where FSRS is enabled, locate "Desired Retention", enter your target decimal (e.g. 0.88 or 0.90), and click Save.

Do I need to change learning steps when using FSRS?

Yes. Under FSRS, learning steps should be short and intra-day (e.g. "10m" or "15m"). Multi-day learning steps like "1d 3d" should be avoided because FSRS calculates optimal inter-day intervals starting from Day 1.

Anki FSRS Calculator: Target Retention & Review Time

Your desired retention is probably wrong, and it's costing you hundreds of reviews a week. Calibrate your target retention to see the exact trade-off between recall and daily review burden.

What Should My FSRS Target Retention Be?

For most learners, the optimal Anki FSRS target retention is 0.88 to 0.90 (88%–90%), with 0.90 serving as the official default. This target balances maximum board exam recall against daily study hours. Lowering retention to 0.84–0.85 during intense clinical rotations cuts review volume by 30%–35%, while rates above 0.90 trigger exponential review overload.

Select Exam or Study Scenario Preset

Simulation Parameters

FSRS-4.5 / 5.0 Math

Desired Retention (Target)90%

80% (Extreme Speed)88-90% (Sweet Spot)95% (Max Recall)

New Cards Per Day25 cards

Active Deck Size (Total Cards In Circulation)15,000 cards

Average Speed Per Card8 seconds

Calculated Review Workload

350reviews / day

Daily Study Time

47 min/day

(5.5 hrs/week)

Time Delta vs 90%

+0h

0% fewer reviews

Cognitive Efficiency Verdict

Optimal Board Balance: 90% retention is the evidence-based sweet spot for USMLE exams. You maintain strong diagnostic recall without hitting exponential review fatigue.

Target Retention Trade-Off Matrix

See how your daily workload changes as you adjust your target retention under FSRS:

Target Retention	Daily Reviews	Daily Study Time	Monthly Commitment	Recommendation
80%	165 reviews	22 min/day	11 hrs/mo	Extreme triage / short rotations
85%	227 reviews	30 min/day	15 hrs/mo	Clerkship sweet spot (30% time saved)
90%	350 reviews	47 min/day	23.5 hrs/mo	Gold standard for Step 1 & Step 2 CK
95%	719 reviews	96 min/day	48 hrs/mo	Diminishing returns / severe review overload

Ready to audit your actual deck?

Upload your .apkg or sync Anki Desktop in 1-click to measure your True Retention and find your failure patterns.

Understanding the Retention vs. Review Volume Curve

In cognitive psychology and spaced repetition mathematics, the relationship between target retrievability and review frequency is non-linear. In the FSRS algorithm, inter-repetition intervals scale inversely with target retrievability (I ∝ −ln(R)), meaning that as you demand near-perfect recall (>90%), review intervals must contract exponentially.

Dialing desired retention back from 90% to 85% yields an estimated 30% to 35% reduction in daily review volume based on FSRS interval formulas. In large-scale empirical benchmarks across millions of review logs, FSRS achieves equivalent retention with 20% to 30% fewer reviews than legacy SM-2 (Ye et al., 2022). This workload dividend reclaims 15–20+ hours per month for clinical question banks (UWorld, Amboss) and high-yield problem solving.

For the full settings walkthrough, including learning steps, relearning steps, and when to press Optimize on your FSRS parameters, see our FSRS vs SM-2 settings guide.

Frequently Asked Questions

Primary Literature & System Citations

FSRS Algorithm & Empirical Benchmark: Ye J, Su J, Cao Y. A Stochastic Shortest Path Algorithm for Optimizing Spaced Repetition Scheduling. Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD '22). 2022:4381-4390. doi:10.1145/3534678.3539081.[ACM Digital Library][arXiv:2202.04600]
Spaced Repetition & Forgetting Index Foundations: Wo&zacute;niak PA, Gorzela&nacute;czyk EJ. Optimization of repetition spacing in the acquisition of knowledge. Acta Neurobiol Exp (Wars). 1994;54(1):59-62.[PubMed PMID: 7943024]
Official FSRS Implementation & Default Specifications: Ankitects. Deck Options: Free Spaced Repetition Scheduler (FSRS). Anki Manual. Published 2023; updated 2024.[Anki Manual Documentation]
Workload-Retention Trade-off & Optimal Retention Derivation: Ye J. The Optimal Retention. FSRS4Anki Wiki. Open Spaced Repetition.[GitHub Wiki Reference]

Clinically & methodologically reviewed by Jed Whetten, MD Candidate (Univ. of Arizona COM-Phoenix)Published May 2024 • Last updated October 2026