🔨 ClozeForge
COGNITIVE RETENTION & MEMORY SCIENCE

The Mathematics of
Never Forgetting.

For 37 years, spaced repetition software relied on Piotr Woźniak’s SM-2 algorithm from 1987. FSRS-5 revolutionizes cognitive acquisition by modeling memory as a 3-dimensional dynamic system: Difficulty, Stability, and Retrievability.

EMPIRICAL DECAY FUNCTION

Forgetting Curves: FSRS-5 vs SM-2

FSRS-5 (Optimal DSR) SM-2 (Static 1987)
Optimal Recall Boundary (90% Desired Retrievability)SM-2: 40% premature over-reviewsFSRS-5: Target Scheduled (Day 28)100%90%50%0%Day 0Day 7Day 14Day 28
Figure 1: Comparison of memory decay modeling. SM-2 forces linear intervals leading to either premature review fatigue or catastrophic forgetting. FSRS-5 precisely targets the 90% retrievability threshold.
THE DSR FRAMEWORK

The Three Variables of Memory

In traditional flashcard systems, a card is simply 'due' or 'not due'. FSRS-5 models every sentence as a continuous point in 3-dimensional memory space:

VARIABLE 1
D • Difficulty

Represents the intrinsic complexity of the linguistic item. Measured on a continuous scale from 1 (trivial cognate) to 10 (intense cognitive load, e.g. irregular subjunctive exceptions).

Range: [1.0, 10.0]
VARIABLE 2
S • Stability

The number of days required for recall probability to drop from 100% to your desired threshold (90%). As stability grows, review intervals expand exponentially without forgetting risk.

Unit: Days (e.g. S = 42.5 days)
VARIABLE 3
R • Retrievability

The instantaneous probability that you will successfully recall the target sentence blank at any given moment t days since your last review.

Range: [0.0%, 100.0%]
MATHEMATICAL SPECIFICATION

Core Mathematical Formulas of FSRS-5

The mathematical engine operates via state transition equations across four user feedback ratings: Again (G=1), Hard (G=2), Good (G=3), Easy (G=4).

// 1. Retrievability Forgetting Curve
R(t, S) = (1 + Factor * (t / S)) ^ (-w₁₆)

Where Factor is normalized such that R(S, S) = 0.90. As elapsed time $t$ increases, recall probability drops following a generalized power forgetting curve.

// 2. Initial Stability & Difficulty (First Review)
S₀(G) = w_{G-1}
D₀(G) = w₄ - e^(w₅ * (G - 1)) + 1

Initial stability is directly parametrized by the first grade. Initial difficulty is inversely bounded so that 'Easy' ratings immediately lower cognitive friction.

// 3. Stability Growth on Successful Recall (Good / Easy)
S'_r(D, S, R, G) = S * (1 + e^w₈ * (11 - D) * S^(-w₉) * (e^(w₁₀ * (1 - R)) - 1) * h(G))

Notice the term e^(w₁₀(1-R)) - 1: when you review a card at very low retrievability R and still get it right ('desirable difficulty'), stability receives a massive cognitive bonus.

// 4. Stability Decay on Lapse / Forgetting (Again)
S'_f(D, S, R) = w₁₁ * D^(-w₁₂) * ((S + 1)^w₁₃ - 1) * e^(w₁₄ * (1 - R))

Unlike SM-2 which mercilessly resets intervals back to 1 day on any mistake, FSRS-5 preserves a fraction of your previously acquired neural stability trace.

PARAMETER VAULT

Decoding the 17 Weights of FSRS-5

These weights are trained on over 200,000,000 empirical review logs using maximum likelihood estimation (MLE).

WeightParameter NameMathematical RoleDefault Value
w₀ - w₃Initial Stability S₀(G)Determines initial stability after 1st review for grades: Again (1), Hard (2), Good (3), Easy (4).0.40, 1.18, 3.17, 15.69
w₄Initial Difficulty D₀Baseline difficulty anchor for new items before user rating adjustments.7.21
w₅Difficulty Grade ScalingStep size for difficulty adjustment based on first response grade.0.53
w₆Difficulty DeltaLinear rate of difficulty adjustment on subsequent reviews.1.10
w₇Mean Reversion WeightDamps extreme difficulty shifts by pulling Difficulty back toward the global mean.0.15
w₈Stability Base Recall GainExponential scaling factor for stability increase upon successful recall.1.65
w₉Difficulty Damping on StabilityControls how much high card difficulty reduces stability growth.0.14
w₁₀Retrievability Decay SensitivityHigher values reward successful recall when retrievability R was low (desirable difficulty).0.94
w₁₁Post-Lapse Stability BaseInitial stability floor assigned immediately after a card lapse (forgetting).2.18
w₁₂Lapse Difficulty ScalingHow much card difficulty penalizes recovery stability after a lapse.0.06
w₁₃Prior Stability PreservationFraction of pre-lapse stability preserved after a forgotten review.0.33
w₁₄Lapse Retrievability DampingImpact of retrievability at lapse time on post-lapse stability.0.28
w₁₅Hard Grade PenaltyStability growth multiplier when user selects 'Hard' (G=2).0.22
w₁₆Easy Grade BonusStability growth multiplier when user selects 'Easy' (G=4).2.94
HISTORICAL BENCHMARK

Evolution of Spaced Repetition (1972 → 2024)

A head-to-head comparison across 5 generations of cognitive scheduling algorithms.

AlgorithmMemory ModelScheduling Formula90-Day RetentionWorkload Efficiency
Leitner System (1972)1D (Box index)Fixed interval doubling~65%High (Rigid buckets)
SuperMemo SM-2 (1987)1D (Ease Factor)I(n) = I(n-1) * EFBaselineFixed intervals
Duolingo HLR (2016)2D (Half-Life)Logistic regression on featuresCloud-reportedAdaptive
FSRS-4.5 (2023)3D (DSR Model)17-parameter power forgettingOpen benchmarkFewer than SM-2 (design goal)
FSRS-5 (2024, open source)3D (DSR + Mean Reversion)Refined power law with lapse decayOpen benchmarkFewer than SM-2 (design goal)
ACADEMIC CITATIONS & RESEARCH

Scientific Literature & Empirical Studies

1. Ye, Jarrett. (2024). Free Spaced Repetition Scheduler (FSRS): A Modern DSR Memory Model. Open-Source Cognitive Science Consortium.
2. Woźniak, Piotr. (1990). Optimization of repetition spacing in computer-assisted learning (SuperMemo 2). University of Technology in Poznań.
3. Settles, Burr & Meeder, Brendan. (2016). A Trainable Spaced Repetition Model for Language Learning. Proceedings of ACL, pages 1848–1858.
4. Ebbinghaus, Hermann. (1885). Über das Gedächtnis: Untersuchungen zur experimentellen Psychologie. Duncker & Humblot.
FAQ

Frequently Asked Scientific Questions

Q: Why is 90% retention the recommended default?

Targeting 95%+ retention requires an exponential surge in daily reviews (roughly doubling review load for just a 5% gain). Targeting 85% causes too many lapses. 90% sits at the exact mathematical sweet spot of knowledge acquisition efficiency.

Q: Can I customize my target retention in ClozeForge?

Yes! In Web App 2.0 settings, you can tune your target retention between 80% (light maintenance) and 97% (high-stakes exam prep) and watch review intervals adjust instantly.

APPLIED COGNITIVE SCIENCE

Experience FSRS-5 in Live Practice

Stop wasting 30% of your time on premature reviews. Train with the 2024 memory engine in Web App 2.0.