Evolutionary updating shapes the predictability of strategy selection in repeated games: code and data

This record contains the simulation code, simulation output and figure scripts for the paper "Evolutionary updating shapes the predictability of strategy selection in repeated games" by Shahariar Chowdhory, Md. Kamruzzaman and K. M. Ariful Kabir, submitted to Chaos, Solitons & Fractals. The study looks at how consistently evolution selects the same strategy when all 32 deterministic memory-one strategies compete in repeated games with execution errors. The games are the Prisoner's Dilemma, Chicken, Stag Hunt and Harmony. Stationary payoffs come from an exact four-state Markov chain. Evolution is modelled with well-mixed replicator dynamics and with death–birth and birth–death updating on random graphs, square lattices, scale-free and small-world networks. Contents- simulation_code/: the code that defines the strategy space, computes noisy pairwise payoffs, builds the networks and runs the population dynamics, plus one runner script for each analysis in the paper. A README maps each runner to the data file it writes.- data/: the final output of every analysis as 46 NumPy (.npz) files. DATA_INDEX.md lists every array in every file, with its shape and type.- figure_code/: Python scripts that reproduce all 15 figures of the paper (Figs. 1–10 and A1–A5).- source_figures/: vector PDF plots made directly from the data. The figure scripts read their values back from these files at full precision. Requirements: Python 3.12, NumPy and Matplotlib. The figure scripts also need PyMuPDF. License: code under the MIT License, data under CC BY 4.0.