Editing Openai/693b7cce-38f4-800d-92f7-5e56a467bfa7 (section)

=== Suggested initial parameter ranges (tunable): ===
* α∈[1.0,2.0]\alpha \in [1.0, 2.0]α∈[1.0,2.0] — degree attenuation exponent. Lower = longer reach.
* M∈[0.01,0.2]M\in[0.01, 0.2]M∈[0.01,0.2] — mutation rate. Small for factual broadcasts, larger for juicy rumors.
* λi∈[0.01,0.2]\lambda_i \in [0.01, 0.2]λi∈[0.01,0.2] (per minute when modeling real time) — memory decay. Higher for low-salience facts.
* ρi∈[0.1,1.0]\rho_i \in [0.1, 1.0]ρi∈[0.1,1.0] — arousal decay.
* κ∈[0.5,5.0]\kappa \in [0.5, 5.0]κ∈[0.5,5.0] — how much surprise drives arousal.
* β\betaβ coefficients in sharing prob: choose so baseline PPP is small (0.01–0.1) and rises under high EEE/AAA/NNN as appropriate. E.g. β0=−4,βE=3,βA=2,βC=3,βN=4,βT=3\beta_0=-4, \beta_E=3, \beta_A=2, \beta_C=3, \beta_N=4, \beta_T=3β0=−4,βE=3,βA=2,βC=3,βN=4,βT=3. Tune to desired behavior.

Use smaller Δt\Delta tΔt for stability if using discrete updates (e.g., game tick = 0.5–1s).