Screenshot + manifest: turning blurry chart captures into structured evidence

Mon, 02 Mar 2026 22:00:00 +0900

1k_scanner is not a document scanner. It is a Rust+egui based multi-market, multi-timeframe trading scanning app.

Most review failures look similar:

You kept screenshots, but forgot why they mattered.
The decision felt clear in real time, then looked vague later.
The same mistake repeated in the next session.

So this post defines screenshot + manifest as a visual style choice and, more importantly, as a decision-evidence routine.

1) Why screenshots alone become blurry evidence

Screenshots preserve scenes, not interpretation.

When the market moves fast, people often lose:

why a symbol was kept as a candidate,
how EMA/NRZ was interpreted at that moment,
what the next action was supposed to be.

That is why a manifest should stay simple: a short interpretation card attached to each screenshot.

2) Lock the visual model first: row=timeframe, column=symbol

In 1k_scanner, keep this model fixed:

Rows = timeframes
Columns = symbols

Following one column keeps one symbol’s MTF context intact. Moving across columns makes candidate comparison faster.

Then switch view modes quickly:

Cmd/Ctrl + 7: dense grid (broad scan)
Cmd/Ctrl + 8: expanded grid (focused review)
Space: single-chart focus toggle

The point is not “see more.” It is to repeat scan wide → validate deep.

3) Use EMA/NRZ as observation sentences, not formulas

At user level, EMA/NRZ can be reduced to two checks:

Is persistence still visible on EMA?
Around NRZ, does structure hold or break?

Do not force certainty. Record what you observed in one sentence.

Example notes:

“EMA persistence is intact, but waiting for cleaner NRZ re-entry.”
“Reaction above NRZ is alive, keep this candidate.”

Those short lines keep continuity in the next session.

4) Read consensus as a priority cue only

In 1k_scanner, when multiple signals align, you see green/red frame emphasis around charts.

For users, this means:

not an execution button,
but a fast way to narrow which candidates deserve immediate review.

Practical use:

stronger frame emphasis → review sooner,
weak/neutral emphasis → defer or drop faster.

So consensus should be treated as an attention-allocation tool, not a prediction shortcut.

5) Make screenshot + manifest one set with `V` and `N`

Speed matters in live sessions.

V: add/toggle current chart in Check Note
N: open Check Note list for batch review

A compact 3-line manifest is enough:

Why this became a candidate (context)
One-line EMA/NRZ interpretation
Next action (continue watching / validate on condition / exclude)

This turns screenshots from “images” into “evidence.”

6) Use templates to preserve repeatability

If you want consistency, lock the routine with templates:

Set Grid Size, Timeframes per row, Exchange, then run Generate Grid Template
Save session layout with F12
Reload next session with Cmd/Ctrl + L

A simple two-template setup works well:

scan template: broad candidate collection,
focus template: deep validation on a short list.

This removes setup drift between sessions.

7) Practical 12-minute loop

0:00–2:00 Load template (Cmd/Ctrl+L), scan in dense mode (Cmd/Ctrl+7)
2:00–4:00 Narrow candidates using consensus frame emphasis
4:00–7:00 Validate candidates in expanded/single modes (Cmd/Ctrl+8, Space)
7:00–10:00 Mark with V, review/edit manifest notes with N
10:00–12:00 Save next-session state with F12 and finalize keep/drop list

The goal is simple:

not more captures, but faster repeatable decisions under the same criteria.

One line to keep:

Screenshots preserve scenes. Manifests preserve decisions.

Use both together, and your next-session judgment gets faster.

Manifest on 1K Scanner — Official Blog