Historical Versus Recent Stability of Bullish Engulfing Intraday Boundaries in XAUUSD M15

This study evaluated whether historically selected contiguous intraday boundaries for unfiltered Bullish Engulfing events in XAUUSD M15 remained viable in recent market conditions. The evidence universe comprised 5,639 events observed from 24 April 2023 through 1 July 2026 and 50 direction-risk-reward configurations, formed from BUY and SELL directions across a five-by-five absolute risk-reward grid. Historical boundary discovery produced 10 direct boundaries, 11 boundaries obtained through a controlled fallback process, and 29 configurations with no qualifying boundary. The historically selected boundary for each qualifying configuration was then frozen and evaluated without reselection over recent 30-day and recent 7-day horizons. The recent 30-day horizon contained activity for all 21 historically qualified boundaries. Only five configurations produced positive cumulative R, while 16 were negative, yielding an aggregate recent result of -251R. Directional divergence was pronounced: all 14 historically qualified BUY boundaries were negative in the recent 30-day window, whereas five of seven SELL boundaries remained positive. The recent 7-day horizon was substantially more sample-limited, with two to four trades per configuration. Three configurations were positive, 16 were negative, and two were flat, for an aggregate result of -97R. Both direct and fallback historical evidence classes exhibited recent weakening, indicating that deterioration was not confined to the less stringent discovery path. The findings show that historical boundary strength and current viability are distinct empirical questions. Historical qualification did not guarantee persistence, and recent performance depended materially on direction and risk-reward geometry. The 30-day horizon provided the clearest current-viability signal, while the 7-day horizon functioned primarily as a high-sensitivity diagnostic because of sparse observations. The results support a frozen-rule monitoring framework in which recently observed outcomes are evaluated against historical boundaries without re-optimisation. They do not establish permanent decay, causal hour effects, or future profitability. Bullish Engulfing events were generated through a proprietary rule-based identification framework, with the exact logic withheld to protect trade-secret methodology.