Two players: Player 1 (Trust-Builder) repeatedly chooses s_t ∈ {0,1} to signal trustworthiness; Player 2 (Trustee) observes public history and a trust state T_t∈[0,1] and chooses a_t ∈ {0,1} to trust or not.
Decision rule: Player 2 uses a threshold θ (a_t = 1 iff T_t ≥ θ). The threshold rule is optimal for Player 2 under Markov strategies.
Equilibrium and sender's optimal policy
Existence: There exists at least one Markov Perfect Equilibrium in pure strategies because state space [0,1] is compact and Bellman operators are contractions.
Sender (Player 1) optimal policy: there is a cutoff \hat{T} such that Player 1 signals (s=1) only when T < \hat{T}; when T≥\hat{T} further signaling is suboptimal. This arises from monotonicity of continuation values.
Two equilibrium regimes: cooperative (trust maintained because signaling is credible) and non-cooperative (trust withheld and signalling costs are sunk).
Simulation and comparative statics — memory λ and threshold θ
Main empirical insight: lower λ (short memory) → trust rises quickly; higher λ (long memory) → slow accumulation of trust; whether cooperation occurs depends jointly on λ and θ.
Illustration: three sample λ values used: 0.3 (fast), 0.6 (medium), 0.9 (slow). For θ=0.5 trust crosses threshold earlier for smaller λ.
Trust placed gradually later (fewer a_t = 1 early)
Table based on simulation summary and textual description.
Heatmap of cooperative frequency (policy implication)
The heatmap shows proportion of rounds with a_t=1 across grid of λ (x-axis) and θ (y-axis): high cooperation region corresponds to low λ and/or low θ; cooperation probability declines when both λ and θ are high.
Practical implication: tailor signaling intensity to audience memory (λ) and required trust threshold (θ) to maximize cooperative outcomes.
报告题目为 “The Trust-Building Game: A Model for Sustainable Cooperation”,作者为 Madjid Eshaghi Gordji 与 Mohamadali Berahman,隶属 Semnan University 的 Faculty of Mathematics, Statistics and Computer Science,日期标注为 December 22, 2025。