When Three Timeframes Agree: The 9-Point Win Rate Edge
M15/H1/H4 directional confluence checking that achieves 64% win rate on aligned signals versus 55% on single-timeframe.
Three Timeframes, One Direction
S28 checks whether the M15 signal direction aligns with the H1 and H4 trend. Alignment is defined as: H1 and H4 both showing momentum in the same direction as the M15 signal, using simple moving average slope and RSI direction as proxies.
When all three timeframes agree, win rate jumps from 55% (M15 alone) to 64% (confluence). That 9-point edge is one of the largest single-feature improvements in the V7 system. Confluence signals also show higher average R per trade (1.12 vs 0.82) because multi-timeframe agreement indicates stronger, more sustained moves.
How Confluence Interacts with L1
S28 is not a separate signal generator. It modifies the L1 signal's confidence score. Full confluence adds +0.05 to calibrated confidence. Partial confluence (2 of 3 timeframes aligned) adds nothing. Counter-confluence (higher timeframes opposing M15) subtracts 0.03 from confidence.
This confidence modification flows through to S14's sizing logic, meaning confluence signals get larger positions. The sizing effect compounds the win rate effect: better signals with larger positions produce disproportionately more R than their frequency suggests.
Why 64% Does Not Apply Everywhere
Full confluence occurs on only about 35% of all signals. If you restricted trading to confluence-only, you would have roughly 1,500 trades over 7.5 years instead of 4,505. The 64% win rate on that smaller sample might not survive because you would be concentrated in trending conditions and would miss the diversification benefit of trading mean-reverting and transitional regimes. S28 adds value through confidence modification, not through hard filtering. It says "this signal is better, bet more" rather than "only trade these signals." That distinction matters for portfolio construction and drawdown characteristics.