Same fill in research and live.
Built so the live seat does not have to invent a different story.
The platform exists to close the gap between a research P&L and a seat. Honest 1-minute fills. Lock-quality data. The same stop path the live software uses. If a result needs a kinder fill model, it is not a result.
Fill model
What “honest 1-minute” means
Default and only acceptable fill for a listed book. Missing 1-minute bars is an error, not a silent fallback to a coarser bar that cannot see the gap.
Entry
The signal bar closes first. The fill is the open of the next 1-minute bar — the first price the seat could have lifted or hit after the decision was knowable.
Stop
Stops walk 1-minute bars. On each 1-minute bar, the stop in force at the open is honored before the trail can move. A gap through that stop fills at the open, not at the skipped price. A gap through the stop fills at that bar’s open, not at the stop price the market skipped.
Costs
Published walks include commission. The desk quote is $0.25 / side ($0.50 / turn) unless a page says otherwise. A number without costs is not quoted.
No silent fallback
Coarser fills are research-only and labeled inaccurate. Old locks from a flatter fill era are invalid until they are walked again on 1-minute.
Reliable data
The tape the lock was fit on is the tape you quote.
If the lock used session volume and aggressor from Tape X, the walk is only valid on Tape X — not a shorter or differently adjusted file.
A history pull with dropped fields or half the volume of the lock tape cannot validate a book. Live analog has to be the same class of data the research used.
When two files disagree on back-adjustment, they are not the same walk. The platform refuses to pretend they are.