Merlin verdict: Markov regime-state method (Lewis Jackson) for GTX — E8 Intelligence Research

**Verdict on the YouTube Markov Regime Strategy (Lewis Jackson)** --- ### (1) What is genuinely sound - **State classification from log-returns** is a valid, simple way to segment regimes. The 20-bar window is arbitrary but not unreasonable for capturing medium-term momentum. - **Transition matrix diagonal dominance** (stickiness) is empirically observed in trending markets (e.g., equity indices during long bull runs). The idea of using *conditional probability* for directional bias is statistically sound as a first-order approximation. - **Walk-forward** is correctly identified as necessary to avoid lookahead bias. The upgrade path to HMM is conceptually correct (though not implemented in the video). --- ### (2) What is flawed or overhyped - **Regime lag is fatal for timing.** A 20-bar lookback means the "bull" state is confirmed *after* the move. The signal enters late, capturing residual momentum at best, but often catching reversals. This is not a prediction—it's a smoothed pas Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com