Workflows

Four scenario cards, all in the same shape. None of them is a recipe. Each is a posture you can adapt to your own reading — the structure is consistent so you can skim diagonally when you return, and so the four cards teach together as well as apart. If any card looks close to your actual workflow, treat it as a starting frame, not a template to copy. If it does not look close, it is still probably teaching something about the two or three cards that do.

Each card carries:

• Setup. The configuration the card applies to.

• Reading routine. How you read the pane during active use.

• Failure mode. The specific way this scenario breaks when it breaks.

• Anti-pattern contrast. The configuration or posture this scenario is explicitly not.

• Expected behavior under stress. What the pane will do when price action gets busy, and how not to confuse that with the tool being wrong.

• What would change this read. A single sentence that names the one condition under which the whole card no longer applies.

Scenario 1 — The differentiated five-slot stack

This is the scenario every STR reader eventually works toward. It is also the one where the duplication trap is most easily avoided if the configuration is deliberate.

Setup. Five slots, all enabled, deliberately differentiated along at least one axis other than timeframe. For example:

• Slot 01 — chart symbol, timeframe 5, RSI length 14, EMA smoothing, EMA signal, On Bar Close? = true, weight 20.

• Slot 02 — chart symbol, timeframe 15, RSI length 9, Jurik smoothing, SMA signal, On Bar Close? = true, weight 20.

• Slot 03 — chart symbol, timeframe 60, RSI length 21, Laguerre smoothing, EMA signal, On Bar Close? = true, weight 20.

• Slot 04 — chart symbol, timeframe 240, RSI length 14, SMA smoothing, SMA signal, On Bar Close? = true, weight 20.

• Slot 05 — chart symbol, timeframe D (daily), RSI length 14, EMA smoothing, EMA signal, On Bar Close? = true, weight 20.

The key discipline is that no two slots are using identical source, length, and MA family. Each slot has a role: a fast intraday read (slot 01), a medium intraday with a responsive smoothing (slot 02), a slow intraday with a low-lag family (slot 03), a multi-session context (slot 04), a session-scale context (slot 05). Weights are equal here for legibility; in production you might lean the weights toward the horizon you actually trade.

Note the slot-05 alignment current behavior — enabling slot 05 disables the all-slot alignment alerts. If your workflow relies on alignment, either hold slot 05 disabled, or treat alignment as a four-slot-maximum signal per the caveat on Alerts.

Reading routine. Walk the four-stage reading order from Visuals and Logic: slot colors first, slot values second, blend color third, structure features fourth. The gain from a differentiated five-slot stack shows up in how often the slots disagree in useful ways — a fast slot flipping before a slower slot, a slow slot holding regime during intraday chop, a cross-family read diverging from a same-length cousin. Those disagreements are the evidence.

Failure mode. The stack degenerates into duplication. Two slots end up with the same source, length, and MA family, and their contributions to the blend stop being independent. The visible result is several slots lighting up together during every regime change. The invisible result is that alignment reads that felt informative were actually redundant.

Anti-pattern contrast. Five slots all at close, RSI length 14, SMA smoothing, SMA signal, across 5/15/60/240/D. The configuration looks thorough; it is one measurement sampled at five cadences. The visual alignment looks persuasive, but it is not breadth. In the current source, the built-in all-slot alignment alerts are also blocked when slot 05 is enabled by the slot-05 counter behavior named in Alerts. Even after that upstream behavior is remediated, this configuration would still be a smoothed average of near-duplicates, not five independent witnesses.

Expected behavior under stress. During hard directional moves, the slots' colors will align quickly — fast slots first, slower slots catching up. During chop, differentiated slots will disagree, and the blend will pull toward the weighted mean of the disagreement. Neither is a flaw; both are the stack reporting what it sees.

What would change this read. If you cannot articulate, in one sentence each, what role every slot plays.

Scenario 2 — Cross-ticker slot as a proxy voice

The cross-ticker slot is STR's most structurally new surface versus Base. Used carefully, it is a cheap way to carry an external voice inside the blend. Used carelessly, it is how readers confuse themselves.

Setup. A four-slot chart-symbol stack (slots 01–03 at default timeframes 5/15/60, slot 04 at 240) plus slot 05 pointed at a related but different instrument. Possibilities: a sector benchmark against a single-name chart, a correlated name against the chart symbol, a volatility proxy, a liquidity-adjacent instrument. Slot 05's weight starts low — a 10% share against 22.5% apiece on the other four is a reasonable first try. The cross-ticker slot is intentionally a minority voice.

Note again that enabling slot 05 disables the all-slot alignment alerts at the time of this pack. If your workflow values alignment, consider running the cross-ticker on slot 04 instead and leaving slot 05 disabled.

Reading routine. Before you start reading the pane with the cross-ticker slot active, read it without. Disable slot 05 temporarily, note the blend and its structure features, then re-enable slot 05 and see how the cross-ticker's contribution shifts the blend. That before-and-after is the main way to feel what the proxy voice is actually doing.

During active use, the cross-ticker slot is a question: is the blend's story consistent with or at odds with the proxy? When the chart-symbol slots are uniformly bullish and the proxy slot is faded, the blend is reporting a tension that would not be visible without the proxy.

Failure mode. The cross-ticker is mismatched. A small-cap single-name chart paired with a large-cap index benchmark is not asking a coherent question — the proxy moves on liquidity and regime dynamics that the chart symbol is not subject to. The blend gets a voice it cannot interpret. Raising the proxy's weight compounds the mismatch.

Anti-pattern contrast. Pairing two instruments that look similar on the chart but move on very different liquidity dynamics, weighting the proxy heavily, and then reading the blend as if the proxy's contribution were comparable to the chart-symbol slots'. It is not.

Expected behavior under stress. During regime changes, the proxy's response will usually lead or lag the chart symbol's, and the blend will show that disagreement as a softer color or a delayed flip. That is the feature, not the bug.

What would change this read. If the proxy and the chart symbol share a driver tightly enough that the proxy stops adding information.

Scenario 3 — Divergence-as-question discipline

This card exists because the divergence triangle is the most magnetic visual on STR and the most easily misread. The posture here is not "how do I use divergence triangles" — it is "how do I avoid turning divergence into an entry."

Setup. Any stack configuration where divergence is on, Plot On Pivot = OFF (the default), and Pivot Len at its default of 20. Keltner can be on or off; the card does not depend on it. The pane does not need any special configuration for this scenario — the discipline is entirely in how you receive the triangle when it prints.

Reading routine. When a divergence triangle prints, do not treat the bar as a decision bar. Instead, answer three questions in sequence, in this order, and out loud or in writing if you can:

1. What is the underlying slot stack doing at this exact moment? Open the slot lines (turn Hide Plot off temporarily) and look. If the stack disagrees internally — some slots full-tone, some faded, the blend not matching the majority — the divergence is floating on top of a mixed reading. That does not invalidate the triangle; it adds a question about which part of the blend is driving it.

2. What happened at the two pivots that produced this triangle? Scrub back Pivot Len bars and look at the original pivot. The triangle describes a geometry between two pivot pairs; the two pivots themselves are the evidence, not the triangle. Triangles can form at pivots that look meaningful and at pivots that look like noise, and the engine does not distinguish between those — you do, by looking.

3. If this triangle were the only input to a decision, would you make the decision? The honest answer is almost always no. The triangle is a question, not an answer. If the answer is yes, you are probably trading the triangle, not the underlying read — and trading the triangle is the posture this scenario is explicitly built to prevent.

Only after answering those three questions does the triangle become an input to a decision you are already constructing elsewhere. The pack is not telling you to ignore triangles. It is telling you to put them in a slower reading queue than their visual authority suggests.

Failure mode. The triangle is treated as a trigger. A reader clicks enter-long on a bullish triangle because "divergence fired." The pane has not made a directional call. It has reported a confirmed geometry between two pivots, and the reader has invented the call out of the visual authority of the triangle.

Anti-pattern contrast. Lowering Pivot Len to 5 to "see more divergences," treating the resulting triangle spam as signal, and acting on every marker. Pivot Len is the strictness dial, not the sensitivity dial, and more triangles at lower strictness are not more evidence — they are more markers representing looser pivots.

Expected behavior under stress. During choppy sessions, confirmed divergences can be rare — the pivot engine is strict by default, which is the cost of stricter pivots. The rarity is the feature working. During trending sessions, the cross-pivot comparisons may produce triangles that disagree with the trend in blend-space, which is exactly what the engine is built to report. A bearish triangle in a rising market is not the engine being wrong; it is the engine noticing that blend-space and price-space are describing two different pictures at the two pivots in question.

What would change this read. If Pivot Len has been lowered below its default to chase more triangles.

Scenario 4 — Co-movement inspection drill

This is less a workflow and more a drill you run once and return to occasionally. It is the cleanest way to internalize that the four structure features are not independent reads.

Setup. Any stack configuration where all four structure features are on (Keltner, BBWP, Donchian, divergence). Defaults are fine. You will toggle slot settings during the drill, so pick a paper chart or a time when you are not trading.

Reading routine. Do this in sequence:

1. With the pane running on defaults, read the four structure features' current states. Note: which direction is Donchian pressing, where is the Keltner stretch, is BBWP tall or short, is there a recent divergence triangle.

2. Change one slot's weight by a meaningful amount — from 33.3 to 10, say, or vice versa. Watch every one of the four features move together as the blend's shape shifts. Not one of them stayed still.

3. Disable one contributing slot entirely. Watch the same thing — all four features re-rank immediately.

4. Re-enable the slot. Settle. Now change the slot's MA family (RSI type). Watch again. All four features move together.

That is the construction-level dependency on display. The four features are not agreeing witnesses; they are four framings of the same blend line. Whatever changes the blend changes all of them.

Failure mode. Never running the drill. A reader who has not felt the features' dependency reads apparent confluence as independent confirmation and weights it too heavily.

Anti-pattern contrast. Describing feature agreement with the word "confirms." A divergence triangle does not confirm a Keltner touch. A Donchian press does not confirm a BBWP reading. They co-exist; they share a source; they cannot independently confirm.

Expected behavior under stress. Under strong directional moves, all four features will often swing together — loudly and in sync. The drill teaches that the co-movement is expected, not new evidence.

What would change this read. If the features diverged among themselves during a blend change. That does not happen; if it does, something is wrong with the configuration.

Cross-cutting habits

A handful of habits generalize across all four scenarios and are worth naming once at the end rather than repeating on every card.

• Read before you touch. Almost every configuration misstep starts with the reader touching inputs before they understand what the defaults were already reporting. Run the four-stage reading order first, every session. The pane is usually telling you something before you start making it tell you something else.

• Name every slot's role. If you cannot complete the sentence "slot 03 is here to answer X" for every enabled slot, you have at least one slot that is not earning its place. Rolless slots do not make the blend richer; they make it redundant.

• Structure features last, in order. Donchian, Keltner, BBWP, divergence. The order matches how indirect each feature's read is, and reading in that order prevents the most indirect feature — the divergence triangle — from setting the frame for the rest of the session.

• Alerts are state observations. Wire them into review queues, not into automated executors. If you need transition firing, do the edge detection at the receiving side. The alert stream serves review workflows by default; it serves execution workflows with a thin layer of routing code you control.

• Respect the slot-05 alignment current behavior. If alignment matters to your workflow, keep slot 05 disabled or treat alignment as a four-slot signal until the upstream behavior is remediated. Do not wire something that cannot fire.

• When in doubt, step back. If you find yourself changing settings without being able to say what problem you are solving, close the indicator dialog and let the pane report at its current state for a few minutes. The discipline at the core of this tool is that the pane is an input, not an oracle; configuration changes should serve a reading, not substitute for one.

Where to go next

If a scenario walked you into a tradeoff you want to inspect in detail, Limitations and Trust Boundaries carries the heaviest misread-prevention surface in the pack. If a scenario pointed you at an input you want to revisit, Settings walks every knob with a cost stated in both directions.