This section translates the mathematical skeleton of Sections III into operational simulation scenarios that policymakers and analysts can run to gauge plausible near-term futures. The aim is not to issue deterministic forecasts but to map probability-weighted trajectories under transparent assumptions, identify early divergences, and connect those divergences to actionable levers.
1. Simulation setup (operational template)
To produce a baseline ensemble, we recommend the following practical setup that can be run with the SDE model described earlier:
State initialisation (plausible Indonesian priors):
P0=0.25P_0 = 0.25 (moderate unrest intensity)
T0=0.45T_0 = 0.45 (trust below midpoint)
E0=0.60E_0 = 0.60 (elevated economic stress)
H0=0.10H_0 = 0.10 (latent but nascent black-horse potential)
R0=0.40R_0 = 0.40 (moderate resilience)
Key policy-response priors:
Relief responsiveness U\rho_U baseline: 1.5 (days-to-weeks operational tempo)
Coercion responsiveness K\rho_{\mathcal{K}} baseline: 1.2
Set r=K/U0.8r = \rho_{\mathcal{K}}/\rho_U \approx 0.8 as a neutral starting point (accommodation slightly favored).
Noise model:
Continuous noise with baseline amplitude cont=0.03\sigma_{\text{cont}} = 0.03 on normalized variables;
Rare jumps (Poisson rate J=0.02\lambda_J = 0.02 per week) with moderate jump magnitude (drawn from a fat-tailed distribution) to capture scandal/incident risk.
Calibration targets: map T,E,P,HT,E,P,H to observed indices (trust polls, CPI food inflation, protest event counts, social media attention), normalise to operational ranges, and fit short-term drift parameters using the last 90 days of data.
Ensemble generation: run N=2,000N=2{,}000 stochastic trajectories for 180 days; compute metrics at 7/30/90/180 day horizons.
