feat: MSS swing detection on HTOB (H1) instead of CTOB (M15)

[cfournel]

GetLastSwingHigh/Low in checkForMSSBefore uses M15 bars. Try H1 for more meaningful swing structure detection per ICT.

[cfournel]

Result: H1 swing detection = 0 trades (fc_mss drops 13→2). Root cause: H1 candle bodies are much wider than M15, making the firstLowLevel < lastBodyLevel quality check fail nearly every time. The strict body-level comparison that works cleanly on M15 becomes almost impossible to pass on H1. Reverted. M15 remains the correct timeframe for checkForMSSBefore() swing detection.