Your answer:
Session accuracy:
0 / 0 (—)
Advanced options
Sample size T: drawn uniformly at random in [40, 120] each round.
Spec: Stationary pool ARMA(p,q) with at least one AR root just outside the unit circle (near-boundary);
positive real roots favored. Non-stationary pool ARIMA(p,1,q).
Optional drift α so that Φ(L)Yₜ = α + Θ(L)εₜ (or Φ(L)ΔYₜ = α + Θ(L)εₜ when d=1).
ADF uses intercept (no trend) with lag selection by AIC up to Schwert cap.
Levels yt
First differences Δyt
ACF (levels)
ACF (first differences)
PACF (levels)
PACF (first differences)
ADF test (intercept only)
Δy_t = α + ρ y_{t-1} + Σ γ_i Δy_{t-i} + ε_t
Variance ratio
Truth (revealed after answer)