Workflows

Three concrete ways to use STR, each in the same card shape so you can skim diagonally. None of these are the "right" workflow. They are three reading patterns that the pane supports well, each with an honest failure mode named next to it. Pick one as a starting point, run it for a week, then adapt.

Each card follows the same structure: setup, reading routine, failure mode, anti-pattern contrast, what happens under stress. Read the one you think fits first, then skim the other two so you know when to move.

Workflow 1 — Differentiated five-slot stack, all four structure features active

Who this is for. A reader who wants the widest-reading configuration STR supports — five deliberate voices across timeframes or symbols, with all four structure features drawing on the blend to provide framing. Chart space is real estate you are willing to spend.

Setup.

• Five enabled slots. Differentiated deliberately, not by accident:

• Slot 01: 5m EMA 12/26/9. Source: close. On Bar Close ON.

• Slot 02: 15m EMA 12/26/9. Source: hl2 (slightly less whippy). On Bar Close ON.

• Slot 03: 60m EMA 24/52/9 (wider pair for the HTF). Source: close. On Bar Close ON.

• Slot 04: 240m Jurik 24/52/9 (different family for a different character). Source: close. On Bar Close ON. Open the Jurik K Power User block only if you have a specific reason.

• Slot 05: Optional Ticker set to a correlated benchmark (e.g., SPY if trading single names, or BTCUSD if trading alt pairs). Timeframe 60m. Source: close. On Bar Close ON.

• Weights: 15 / 20 / 25 / 25 / 15. Equal would be defensible; graduated weights match the intent of reading HTF evidence more heavily than the reactive LTF slot.

• All four structure features ON. Pivot Len 20 on divergence. Keltner default multiplier 2.0. BBWP threshold 50. Donchian length 20.

• Hide Plot ON for slots 01 through 04. Hide Plot OFF for slot 05 so the benchmark voice is visible.

Reading routine (daily).

1. Glance at the histogram first. Is the blend above or below 50? Is the column bright or faded? That is your regime-and-thrust read in one look.

2. Scan the slot 05 (benchmark) line. Is it pulling the blend or confirming the blend? If it is pulling, the blend's move is partly a benchmark story — useful context.

3. Donchian and BBWP together. A Donchian upper touch with BBWP expanding is the blend printing a fresh high in an expanding width regime. A Donchian upper touch with BBWP compressed is the blend printing a fresh high in a quiet width regime — different posture, same touch.

4. Keltner only if you see a noticeable stretch. The envelope is drawn to help you notice when the blend is far from its basis; most bars, the blend is not particularly stretched and the envelope is background.

5. Divergence triangles when they print. Check whether the slot stack agrees (all enabled slots with K on the same side as the triangle's implied direction) or argues (slots dispersed). Triangle plus agreement is context. Triangle plus disagreement is a geometry note inside a sustained move.

Failure mode. Chart clutter. Five visible-or-hidden slots plus four structure features on one pane can crowd a small window. If you run this on a 15-inch laptop, your reading time per bar goes up, and the cost of the configuration starts to show. Run it on a monitor that gives the pane a full third of the screen.

Anti-pattern contrast. Five slots at identical EMA 12/26/9 with no differentiation. That is the alignment trap with extra slots attached. The visual alignment read will look like more evidence while reporting almost nothing new. It is exactly the same evidence, sampled five times. Current source note: the built-in all-slots alert pair should not be used as the proof surface for this five-slot workflow while slot 05 is enabled, because the slot-05 active-count path is overcounted.

What happens under stress. When a session gets fast, the histogram and slot 05 are the two reads you fall back to. Everything else drops to background. Do not try to read all four structure features under time pressure — use the routine's first two steps and come back to the rest when the session calms.

Workflow 2 — Divergence-as-question on a quieter configuration

Who this is for. A reader whose process already has entry and exit reasoning and wants STR specifically for the divergence module as a question-asking tool. The blend and the structure features are supporting cast; divergence is the reason this configuration exists.

Setup.

• Three enabled slots at defaults (5m, 15m, 60m, EMA 12/26/9). Hide Plot ON for all three. Equal weights.

• Slots 04 and 05 disabled.

• Divergence ON. Pivot Len 20. Plot On Pivot OFF (start here — see the honesty beat note below).

• BBWP OFF. Donchian OFF. Keltner OFF.

• Master smoothing OFF.

• Overbought / Oversold lines at default 70 / 30 (visual only).

Reading routine.

1. During the session, you are not watching the pane continuously. You are watching price. STR is a listener.

2. When a divergence alert fires, you look at the pane.

3. Check the blend position (above or below 50, close to or far from the fill's color change).

4. Check the histogram state. Is it bright or faded on the side that matches the triangle's implied direction?

5. Un-hide the slots temporarily. Are the slots dispersed (some above their signals, some below) or aligned on one side? Dispersion suggests the blend is averaging disagreement; alignment suggests the slots are telling a consistent story.

6. Return to your entry-and-exit framework with the divergence moment as one input among several. The triangle is the prompt to ask the question, not the answer.

Failure mode. Two versions, in order of how often they actually bite.

The common one: treating the triangle as an entry. You see the triangle, the pane looks clean, your hand is already near the order entry, and the rest of the reading routine quietly collapses. This is the failure mode the pack names on every page that mentions divergence. It shows up here because it shows up everywhere.

The subtle one: confirmation-chasing. You have a view before the triangle prints. The triangle happens to agree with your view. You treat the agreement as confirmation and size up. STR's divergence module does not confirm; it prints a geometry about two confirmed pivots. If you went long because you already wanted to go long and the triangle happened to print, that is a coincidence dressed as evidence. The tool will look like it confirmed the idea; the tool did not confirm anything — it reported a geometric fact, and you imported your existing conviction into the interpretation. Catching this failure mode requires brutal honesty about the order of events. Did you decide to go long, then see the triangle? Or did the triangle genuinely change a view you were still weighing? The first version is the version that costs people.

Anti-pattern contrast. Plot On Pivot ON and Pivot Len 5. More triangles, back-shifted visually, printing on loose pivots. Every triangle you see visually lands on a bar where it was not actually visible in real time, and many of the triangles are low-quality pivots anyway. This configuration produces an impressive-looking chart in the post-hoc review and does not help you in the session.

What happens under stress. The divergence module is slow by design. Pivot Len 20 means confirmation arrives 20 bars after the actual pivot. In a fast session, triangles will land well after the move they describe. That is not a failure of the module — that is strict pivots being strict. If you need faster divergence, use the module in its faster configuration (e.g., Pivot Len 10) and accept that the triangles will be noisier. STR does not offer faster-and-stricter simultaneously.

Workflow 3 — Minimal-configuration-first, learning the pane

Who this is for. A reader new to STR, or a reader who wants to build reading intuition before adding complexity. The configuration is intentionally small.

Setup.

• Three enabled slots at defaults (5m, 15m, 60m, EMA 12/26/9). Hide Plot ON. Equal weights.

• Slots 04 and 05 disabled.

• Plot Blend ON. Plot Histogram ON.

• One structure feature ON — pick Donchian or BBWP. Turn the other three OFF.

• Master smoothing OFF.

• Default Pivot Len 20 (if you chose to keep divergence on instead, which is a valid alternative).

Reading routine.

1. Histogram first, every time. Build the color-state muscle memory.

2. Blend line position (above or below 50, close to or far from its signal).

3. Your one structure feature. Spend a week noticing what it does around specific price actions. Make notes about what surprised you.

4. Toggle the other structure features on one at a time, ideally one per session day. Notice how each one interacts with the reading you were doing before.

Failure mode. Getting bored and turning everything on because "more features" feels like more value. Each structure feature you add without a learned reading pattern dilutes your attention. Go slower than feels natural.

Anti-pattern contrast. All four structure features ON from day one without a reading routine. You will see a busy pane, you will try to read everything at once, you will not internalize any of the layers, and you will end up with a vaguely uncomfortable feeling of "the tool is too much" — which is a failure of the learning path, not the tool.

What happens under stress. This configuration reads cleanly under time pressure because there is so little to read. Histogram, blend, one structure feature. That is the full pane. Fast sessions are easier here than in Workflow 1.

Cross-cutting habits

Whichever workflow you run, these habits will age well:

• Read the pane in the same order every time. Histogram first, slots second, blend third, structure features in the deliberate order you chose. The order is the muscle memory. A consistent reading order is worth more than a clever reading order — it is what lets your attention specialize into noticing what has changed instead of re-discovering the layers every bar.

• Treat alerts as review triggers, not action triggers. Wire them to a queue you watch, not an executor that trades. The pane's alerts are state descriptors; an executor that takes state as action will over-fire in every sustained regime.

• Keep a misread journal. When a divergence triangle or an alignment alert prints and turns out wrong, write down what the rest of the pane said at that bar — where the blend sat, what the histogram was doing, which slots were on which side. You will notice patterns in the misreads faster than you will notice them in the correct reads, because the misreads are where your attention sharpens. Over weeks this journal is worth more than any configuration change.

• Audit the blend weekly. Open the inputs dialog and check that your slots are still differentiated (if running Workflow 1) or still minimal (if running Workflow 3). Configurations drift — a slot you enabled to test something never got disabled, a weight you tweaked for one session is still tweaked, a Power User parameter someone touched last month is still touched. Weekly audits catch the drift before it quietly reshapes what the pane is telling you.

• Leave Sensitivity at 1.0 until evidence moves it. Sensitivity is the easiest knob to over-tune, because changing it produces a visibly different pane and "visibly different" feels like progress. Check your default quarterly; if a session genuinely convinces you to change it, change it and log the reason. A Sensitivity change without a written reason is a knob you will re-tune three weeks from now.

• Defend each slot in one sentence. If you cannot say what slot 05 is for in a single sentence, slot 05 does not need to be enabled. The same test applies to every structure feature you have toggled on.

• Choose by the job. Base is the focused three-slot read. CTX is the ten-slot context surface when breadth across symbols or timeframes matters most. STR is the five-slot structure surface: it keeps the stronger per-slot control pattern, then spends the extra processing budget on Keltner, BBWP, Donchian, and divergence reads around the blended MACD output. If the pane feels busier than you can read, simplify the STR setup first (Workflow 3) and only change versions when the job itself changes.

Next: Troubleshooting for specific symptom-to-cause reasoning, or Limitations and Trust Boundaries for the estimate boundary if you have not read it yet.