Abstract for: An Open-Source Python Replication of the Loop Impact Method for System Dynamics Model Analysis

Hayward and Boswell (2014) introduced the loop impact method for quantifying feedback loop dominance in system dynamics models, implemented in the proprietary STELLA environment. This limits accessibility and reproducibility for the broader SD community. We present FLIP — Feedback Loop Impact in Python - a complete open-source replication of the algorithm and a pathway towards integration with PySD. We translate the full loop-picker algorithm of Appendix 1 of Hayward and Boswell (2014) into Python using NumPy and SciPy, replicating the fixed-step RK4 integrator, STELLA DERIVN-equivalent central-difference derivatives, and junction rules. The implementation is validated against all four benchmark models from the original paper and released as a documented Jupyter notebook on GitHub. The loop impact algorithm is fully reproducible from the paper's text. The first-order limits-to-growth model matches published transition times within 2%. Residual discrepancies in the yeast and market growth models are traced to undisclosed lookup table data embedded in the original STELLA supplement, illustrating a concrete reproducibility gap in the system dynamics methods literature. Open-source implementation of loop impact analysis removes software cost barriers, enables version control, and integrates with Python scientific ecosystems. FLIP is designed as a foundation for a PySD extension that automates structural dominance analysis for any STELLA or Vensim model, embedding loop impact into reproducible calibration, sensitivity analysis, and participatory modelling workflows.