Multi-agent Bayesian Learning with Adaptive Strategies: Convergence and Stability