bollinger_pullback
Admitted 2026-09-21 by owner override — the replay did NOT clear the bar (n=2, PF 0.00, dd 1%, net -36): admitted on the owner's instruction of 2026-09-21: a live audition on the test account regardless of the replay; the retirement rules decide. Every other rule applies: it retires at 40 closed live trades below PF 1.0 or after 120 days, and the live record decides. The hourly bars are resampled from the IEX 5-minute feed — live and in this replay, so both read one tape — and IEX is a thinner slice of the market than the consolidated tape, so an hourly high or low can miss where the wider market actually traded. Whether a stop or a target was HIT is decided on the consolidated tape instead (N7), because that is what a resting order really fills against; a replay that checked the touch on the thin feed would quietly spare the book every stop-out IEX happened to miss. The book also holds overnight, and a broker's stop and target legs do not execute outside regular hours: it is protected during the session and exposed from 16:00 to 09:30, where a gap fills at the opening print rather than at the level — and a bar that OPENS beyond a level is booked at that open, target first, because that order was already marketable when the bell rang. The chapter enters on a stop order a few pips beyond the pattern bar's extreme; this book enters at the close of the pattern bar, because the runner holds one resting order at a time and a ten-seat book would spend the session queueing. The difference is not neutral: the chapter's stop entry is a filter that only takes the reversals which follow through, so entering at the close takes a strictly larger and adversely-selected set — every extra trade is one the chapter's trigger would have rejected. bollinger_pullback and bollinger_pullback_trail are also confounded: their entries are identical by construction and they share one test account, where one owner per symbol means a contested name goes to whichever book signals first — so the difference between their records is not a measurement of the trail.
Last tick 4:00 pm ET · all systems normal
Subscribebollinger_pullback
In testingOwner overridetest accountAdmitted 2026-09-21 by owner override — the replay did NOT clear the bar (n=2, PF 0.00, dd 1%, net -36): admitted on the owner's instruction of 2026-09-21: a l…
Born 2026-09-21 · no funded cell yet · trades the test account · trading live since 2026-09-21
The owner admitted this strategy to the test account although its replay failed the admission bar — a live audition on real fills. The retirement rules are unchanged.
Admission replay, Sep 21, 2026 — did not clear the bar; admitted by owner override. 2 trades, PF 0.00, max dd 1% of cell, net -36 (2y replay, sip) — did NOT clear the bar (pf(0.00<1.05), trades(2<40)); owner override: admitted on the owner's instruction of 2026-09-21: a live audition on the test account regardless of the replay; the retirement rules decide Admitted on 2 trades, PF 0.00; the replay was re-run on Oct 3, 2026 and the column now reads 1 trades, PF 0.00. Universe: 103 names, ranked as they stand today and carried across the whole replay.
Admission note
Admitted 2026-09-21 by owner override — the replay did NOT clear the bar (n=2, PF 0.00, dd 1%, net -36): admitted on the owner's instruction of 2026-09-21: a live audition on the test account regardless of the replay; the retirement rules decide. Every other rule applies: it retires at 40 closed live trades below PF 1.0 or after 120 days, and the live record decides. The hourly bars are resampled from the IEX 5-minute feed — live and in this replay, so both read one tape — and IEX is a thinner slice of the market than the consolidated tape, so an hourly high or low can miss where the wider market actually traded.
Whether a stop or a target was HIT is decided on the consolidated tape instead (N7), because that is what a resting order really fills against; a replay that checked the touch on the thin feed would quietly spare the book every stop-out IEX happened to miss. The book also holds overnight, and a broker's stop and target legs do not execute outside regular hours: it is protected during the session and exposed from 16:00 to 09:30, where a gap fills at the opening print rather than at the level — and a bar that OPENS beyond a level is booked at that open, target first, because that order was already marketable when the bell rang. The chapter enters on a stop order a few pips beyond the pattern bar's extreme; this book enters at the close of the pattern bar, because the runner holds one resting order at a time and a ten-seat book would spend the session queueing.
The difference is not neutral: the chapter's stop entry is a filter that only takes the reversals which follow through, so entering at the close takes a strictly larger and adversely-selected set — every extra trade is one the chapter's trigger would have rejected. bollinger_pullback and bollinger_pullback_trail are also confounded: their entries are identical by construction and they share one test account, where one owner per symbol means a contested name goes to whichever book signals first — so the difference between their records is not a measurement of the trail.
Test strategies trade a real paper account; funded ones trade the funded account. Both are real fills. This strategy trades the test account.