Take, Pass, Watch — and feedback on a read

Decision capture

One tap freezes the read into a record that does not change afterwards. The signal is re-derived at the moment of the decision, so the record matches what was on screen rather than what the read said when it was first computed.

There is a short Undo window that reverses the entire capture. After it closes the row is permanent — an editable trade journal is a journal that flatters whoever edits it.

The Take ticket

Entry, stop, target and size can all be overridden before committing. Any override re-derives the dollar risk, so the size and the risk figure can never tell two different stories.

The consequence for the daily budget is shown before the commitment rather than after it. A trade that would breach the day's loss cap says so while there is still a decision to make.

The adherence flag

Every take is recorded as either followed or override. An override is a take against a no-trade signal, against fired rule flags, or on a low-grade read. A clean take on a strong grade records as followed.

The flag exists so the performance metrics can later answer a question most journals cannot: whether the trades you took against your own rules did better or worse than the ones you took with them.

Pass and Watch

Both are decisions and both are recorded. They resolve into shadow outcomes against later bars, which is how a good pass gets credited — the alternative is a record in which declining a bad trade is indistinguishable from never having looked.

Watch additionally arms a price alert at the first target.

Feedback is opinion, not measurement

The thumbs control and its four disagreement categories are counted, and shown back as a disagreement rate. That is the whole of what they do.

Feedback never reaches a grade, a confidence figure or the calibration. This is not a current policy that might be revisited at a larger sample size — it is enforced by a build check that bans the import outright, because a grading engine that learns from whether users liked its answers is optimising for agreement rather than accuracy.

Frequently asked

Does my thumbs-down teach the grading engine anything?

No, by design and by build gate. Feedback is opinion rather than measurement: it is counted and shown back as a disagreement rate, and a build check bans it from ever reaching grades, confidence or calibration at any sample size.

What does the override flag on a journal row mean?

That you took a read your rules argued against — a no-trade signal, fired rule flags, or a low-grade read. Clean takes on strong grades record as followed, and the performance metrics later compare the two populations.