Bad Forgetting Curves: Interactive Model Comparison

Time Range: 30 days

Active Model for Parameter Control

R(t) = R₀ × (1 + t/τ)^(-β) + R∞

R₀ (Initial Recall): 0.85

τ (Time Scale): 5.0

β (Decay Exponent): 0.80

R∞ (Asymptotic Floor): 0.15

R(t) = R₀ × exp(-(t/λ)^k) + R∞

R₀ (Initial Recall): 0.85

λ (Scale Parameter): 8.0

k (Shape Parameter): 0.60

R∞ (Asymptotic Floor): 0.15

R(t) = R₀ × [w×exp(-t/τ₁) + (1-w)×exp(-t/τ₂)] + R∞

R₀ (Initial Recall): 0.85

w (Fast Component Weight): 0.70

τ₁ (Fast Time Constant): 1.0

τ₂ (Slow Time Constant): 20.0

R∞ (Asymptotic Floor): 0.15

Realistic: Rapid initial decline that gradually slows down
Suspect: Curves that go up, have inflection points, or drop too steeply late
Power Law: Should show characteristic "fat tail" - slow asymptotic approach
Weibull: Very flexible shape, but k<1 gives the slowing behavior you want
Superposition: Can show multi-phase forgetting (fast + slow components)