Directional Asymmetry in Bullish Engulfing Outcomes: A Comparative Study of BUY and SELL Configurations in XAUUSD M15

This study examined directional asymmetry between matched BUY and SELL configurations for unfiltered Bullish Engulfing events in XAUUSD M15. The analysis used 5,639 events recorded from 24 April 2023 through 1 July 2026 and evaluated 25 matched risk-reward geometries, producing 50 direction-specific configurations. Each BUY configuration was paired with the SELL configuration using the same absolute risk and reward settings. The comparison included boundary presence, discovery strength, selected intraday timing, cumulative R, average R, win rate, day-level and week-level consistency, boundary overlap, and recent 30-day and 7-day evidence. BUY produced a qualifying contiguous intraday boundary in 14 of 25 geometries (56.0%), compared with 7 of 25 geometries (28.0%) for SELL. Six matched geometries produced boundaries in both directions, eight produced a BUY-only boundary, one produced a SELL-only boundary, and ten produced no boundary in either direction. The BUY and SELL boundaries within all six dual-boundary geometries were non-overlapping, indicating that opposite-direction historical opportunities were concentrated in different intraday windows rather than representing simultaneous directional signals. BUY exceeded SELL in historical cumulative R in five of the six dual-boundary pairs, while one pair was tied. SELL boundaries were absent at reward settings of 3R, 4R, and 5R. Historical directional advantage was not consistently preserved in the recent evidence horizons. Several historically BUY-favourable geometries weakened or reversed during the recent 30-day and 7-day windows, while the short 7-day samples remained limited. These findings show that the conventional bullish label did not fully determine empirical directional outcomes. Directional behaviour depended on the interaction among trade direction, absolute risk-reward geometry, and intraday timing. Bullish Engulfing events were generated using a proprietary rule-based identification framework. The exact event-identification logic is withheld to protect trade-secret research methodology. The results are historical and exploratory and should not be interpreted as investment advice, a live-trading recommendation, or evidence of future profitability.