We revisit the Multiplexed Hysteretic Threshold Autoregressive (MHTAR) model introduced by Frydrych and Szewczyk (2014) and provide three contributions absent from the original paper. First, we establish a rigorous measure-theoretic foundation: the extended filtration FMHTAR that encodes the full distribution of outstanding orders. Second, we implement a genuine Preisach-plane agent-based simulator — translating the magnetic-hysteresis order-book analogy into a computational model — and demonstrate that independent Preisach-MHTAR series exhibit unstable rolling correlation (R² ∈ 0, 0. 9) driven by endogenous stop-loss cascade trends, not by a shared latent factor (return-level R² ≈ 0). Third, we apply a corrected block-aggregated Hurst estimator to daily S&P 500 returns 2000–2023 (H ≈ 0. 45, p < 0. 001), implement a nested breakout-entry RTSM strategy pool, and benchmark it against a constant-fraction index exposure at the same average market weight. The combined results strengthen the case that physics-inspired, value-dependent models are a necessary complement — rather than replacement — for classical time-series approaches in quantitative risk management.
Frydrych et al. (Sun,) studied this question.