Fibonacci 61.8% pullback, engine self-test 2026-10-08T21:32:37.392Z
[PASS] fractal swing low and high found at bar 2
[PASS] equal lows or highs make no swing
[PASS] case 1 precondition: the impulse qualifies at hours 22 and 23 (size >= 3 x ATR): 3 x ATR 0.00549, 0.00639
[PASS] case 1: one long trade, filled at the recalculated E (not the first one), target at the extended H: {"rawIn":1.10073,"stop":1.095,"target":1.11,"reason":"target"}
[PASS] case 1: costs (0.15 pip each fill + 0.7 pip commission) in R and pips: R 0.00573 net 1.60035 R 91.70 pips; by hand 1.60035 R 91.70 pips
[PASS] case 1 mirrored: one short trade with the same net R: 1.60035 R
[PASS] case 2: gapped below E, filled at the open 1.1000: 1.1
[PASS] case 3: open at or below L cancels the order, no trade: trades 0, gap kills 1
[PASS] case 4: filled at E and stopped in the same minute at L: stop -1.01745 R
[PASS] case 5: stop and target in one minute, stop taken: stop
[PASS] case 6: stop gapped through, filled at the open 1.0930: 1.093
[PASS] case 7a: position held across a 48-hour gap, target after it: 48.983333333333334 h
[PASS] case 7b: order pending across a 48-hour gap, filled after it at E: 1.10073
[PASS] case 8 precondition: the new impulse (L 1.1020, H 1.1080) is below 3 x ATR at hour 27: 3 x ATR 0.00797
[PASS] case 8: new swing low cancels the pending order, no trade: trades 0, setups new swing,open
[PASS] case 9: long and short touched in one minute, both cancelled
[PASS] case 10: after the target, a second pullback to E does not trade the same impulse
real data: 8326700 minutes, 141042 one-hour bars, 2004-01-01T01:00 .. 2026-09-28T16:00
[PASS] 1-hour bars equal a naive grouping of the real minutes: 141042 bars, 0 differences
[PASS] ATR(14) equals a naive Wilder recomputation: largest difference 5.20e-18
[PASS] swings equal a naive scan: 18683 swing lows, 18475 swing highs, 0 differences
[PASS] FB-L: engine and independent re-implementation agree trade by trade: 1927 vs 1927 trades, 0 differences 
[PASS] FB-L: control machinery at shift 0 with the real order path reproduces the real trades: 1927 vs 1927 trades, 0 differences
[PASS] FB-L: control as specified (levels rebuilt from the shifted bar) runs at shift 0: 1278 trades from 5277 setups
[PASS] FB-LS: engine and independent re-implementation agree trade by trade: 3374 vs 3374 trades, 0 differences 
[PASS] FB-LS: control machinery at shift 0 with the real order path reproduces the real trades: 3374 vs 3374 trades, 0 differences
[PASS] FB-LS: control as specified (levels rebuilt from the shifted bar) runs at shift 0: 2363 trades from 10543 setups

PASSED