#46 — FIFO 2-box trail sweep (X3c_2_5)

Hypothesis — what it tests

Does a wider 10-pip (2-box) trailing stop make the FIFO-Trends P&F edge survivable under real spread and slippage, where the 5-pip 1-box trail was consumed live?

Description

After live FIFO-Trends deployments on account 013 died to 1-bar whipsaws (the 5-pip 1-box trail left only +2.1 pips over GBP_JPY's 2.9-pip median spread), this experiment reran the exact same b=5, r=1 configs with trail_d=2 — a 10-pip trail giving a +7.1-pip effective floor. GBP_JPY_ft2 (35.3 p/d OOS, 47% WR) and USD_JPY_ft2 (32.7 p/d, 46% WR) both passed walk-forward, though OOS p/d dropped roughly 50% from the 1-box results (71.6/68.5) because a wider trail means fewer exits and more drift per trade; EUR_JPY_ft2 failed WF outright at +5.9 p/d. Both passing configs were added to the fx-fifo-paper service (R7 consistency check 7/7 PASS) to watch the trail% exit split before any live redeployment. The Session 087 epilogue was negative: the ft2 variants were the worst paper performers (GBP_JPY_ft2 -274p, USD_JPY_ft2 -156p) and the whole FIFO paper book was stopped at -627.3p/849 trades with the standing instruction not to redeploy FIFO-Trends in any trail form.

Key result

Indicators

Algorithms

Code