Settings

This page walks every input with a cost stated in both directions. There is no "best" setting here, and the pack refuses to invent one. A knob with a good default has a good default because the tradeoff at the default is generally honest for the widest range of first-session readers — not because the default is the answer. If you came here looking for "recommended settings for scalping" or "my favorite parameters," the page will disappoint you on purpose. Your favorite parameters are the ones you will build through use; this page is built to keep that building honest.

Read this page three ways. Read it straight through once so you know what surfaces exist. Return to it the first time you are mid-configuration, using the decision map at the top to jump to the knob that matters right now. Return to it a third time later, after the pane has misbehaved at least once in a way you did not expect — the named-misuse lines on each block usually explain the surprise, and reading them with a specific confusion in mind lands differently than reading them fresh.

The page opens with a decision map so that a reader mid-configuration can jump to the knob they need. The map is followed by a tripwire block — four common first-configuration mistakes that cost the most when left unexamined. After that, inputs are walked in the order they appear in the dialog.

Decision map — what to change, in what order

Configure in this order, roughly, and most of the typical misconfigurations stop happening:

1. Decide what each of your slots is for. Five slots is not the question; five slot roles is. A timeframe context ladder. A benchmark voice. A correlated-instrument voice. A volatility-proxy voice. A chart-symbol read at a slower length. Every slot that does not have a role — that is not answering a question the others do not already answer — is a vote for the duplication trap.

2. Choose repaint posture per slot. On Bar Close? is now per-slot on STR. Mixing true and false postures across slots is legitimate but comes with a cost — see MTF and Repainting — and should be a deliberate choice, not a byproduct of touching one slot and not the others.

3. Tune lengths and MA families only where a slot's role demands it. Changing MA family on every slot is usually a sign that the role question above was skipped.

4. Turn on only the structure features you will actually read. All four are on by default for Donchian, divergence, and BBWP; Keltner is off. You are not required to run all four. A crowded pane is not a better pane.

5. Leave master smoothing off until you have a stated reason. It is calmer and slower. Blend-based alerts fire later. Per-slot alerts are unaffected. See the master-smoothing block below.

6. Touch Power User blocks last, and only when you can finish the sentence "I expect this change to move X because Y." They are earned access, not feature density.

Tripwires that bite on a first configuration

These are the four most common and most costly ways a first-pass configuration goes sideways. None of them looks obviously wrong from inside the settings dialog; all of them look obvious in retrospect.

• Mixed-posture repaint without noticing. Setting On Bar Close? = false on an HTF slot while other slots sit at true produces a blend whose members disagree about repaint. The blend is now a weighted mix of confirmed and live-and-repainting values, and structure features over that blend inherit the mix. They can look jumpy for reasons the feature math did not cause. If your first impulse is to reach for master smoothing because "the pane is too noisy," check your slot postures first; fifty percent of the time the noise came from the blend's own ingredients, not from anything that needs damping.

• Pivot Len reduction to see more divergences. Lowering Pivot Len trades strict pivot logic for triangle spam. The cost is that every loose pivot gets counted, and none of them get counted more strictly. Readers who move Pivot Len from 20 to 5 to "see more signals" are paying by volume and losing discrimination. Pivot Len is a strictness dial. It is not a sensitivity dial.

• Power User block poking. Opening ALMA sigma, Jurik phase, or VAMA volume length without a directional hypothesis is the fastest way to degrade a configuration by accident. These inputs reward users who can finish the sentence "I expect this change to move X because Y." They punish users who are clicking to see what happens. If you cannot state the expected direction of the change before you make it, leave the block closed and come back when you can.

• Master smoothing read as "cleaner equals more reliable." Master smoothing adds lag. A calmer plot is not a truer plot. If the pane was choppy, that chop was often evidence — the blend telling you about disagreements among its members that an average would paper over. Smoothing an informative chop into a clean curve does not give you a more accurate read; it gives you a later one.

Per-slot primary inputs (slots 01 through 05)

Every slot carries the same surface. The differences between slots are fixed colors (teal, aqua, blue, orange, yellow for slots 01 through 05, not user-configurable), default enable state (01, 02, 03 enabled; 04, 05 disabled), default timeframes (5, 15, 60 for the enabled three; empty for the disabled two), and default weight (33.3 for the enabled three; 0.0 for the disabled two). All other defaults match across slots.

Enable

On/off. Controls whether the slot computes and contributes to the blend, and whether the slot's alerts can fire. Off is a genuine kill switch for the slot. Defaults: slots 01–03 enabled, slots 04–05 disabled.

Cost in both directions. More enabled slots means more surface to configure and more potential for the duplication trap if the roles are not differentiated. Fewer enabled slots means a narrower blend.

Hide Plot

On/off. Controls only whether the slot line is drawn. Defaults: slots 01–03 hidden, slots 04–05 visible. The default makes the blend legible without five overlaid slot lines.

Misuse to name. Hide Plot is not a kill switch. An enabled, hidden slot still computes, still contributes to the blend, and still fires its own alerts. Readers who use Hide Plot to declutter and later forget that a hidden slot is active will be confused when the blend moves for reasons the visible slots cannot explain.

Source

Which price series feeds the slot's raw RSI. Defaults to close.

Cost in both directions. Close is standard and what most users mean when they say "RSI." HLC3 or typical price is smoother but can disagree with close-based RSI at swing points. Mixing sources across slots (close on one, HLC3 on another) is sometimes deliberate — it lets the slots see the bar differently — but it should be a chosen thing, not a byproduct.

Optional Ticker

Empty string means the slot inherits the chart symbol. Any valid TradingView symbol overrides the chart for that slot only.

What this unlocks. A benchmark voice (e.g. an index against a single-name chart), a correlated-instrument read, a volatility proxy. The cross-ticker slot is STR's most structurally new surface versus Base.

Misuse to name. Pairing a very large-cap benchmark with a small-cap chart symbol produces a blend that is reporting a hybrid read across two different liquidity regimes. That is not inherently wrong — sometimes it is exactly the question you want to ask — but it deserves to be a deliberate choice rather than a side effect of typing a ticker in.

TimeFrame

Empty inherits the chart timeframe. Any selection must be at or above the chart timeframe.

Guardrail. Setting a slot timeframe lower than the chart timeframe raises a runtime error with the exact text: "<slot label> timeframe cannot be lower than the chart timeframe." The fix is either to raise the slot TF or lower the chart TF. More on this and on chart-TF collapse in MTF and Repainting.

RSI Length

Lookback for the raw RSI calculation. Default 14, minimum 1.

Cost in both directions. Longer lengths produce a calmer raw RSI with more memory and slower reactions. Shorter lengths react faster and are noisier. Below about 5, the raw RSI has almost no memory and the subsequent smoothing pass has little to work with. The default 14 is the classical value for reasons that apply to many symbols and timeframes; it is not a best choice, it is a common one.

RSI Smoothing Length

Length of the MA pass applied to raw RSI. Default 3, minimum 1.

Cost in both directions. Longer smoothing produces a plot that lags the raw RSI more visibly. At length 1, the pass is effectively off — no smoothing. Values much above 10 begin to produce a plot whose timing is noticeably detached from the underlying RSI.

RSI Type

MA family for the RSI smoothing pass, drawn from the Axiom MA library. Options include SMA, EMA, RMA, WMA, HMA, TEMA, DEMA, T3, VWMA, ALMA, KAMA, FRAMA, Jurik, Laguerre, and VAMA, among others. Default SMA.

The families differ in how they respond to short, sharp changes versus persistent moves. Picking a family is a statement about what you want the slot to ignore and what you want it to amplify.

Cross-reference. The family mechanics and the Power User parameters that tune them are taught in the Axiom MA library manual. This page does not re-teach them; it names the cost of the decision and points at the right place for depth.

Signal Length / Signal Type

Second MA pass that turns the slot's smoothed RSI into its signal. Defaults: length 3, type SMA, minimum length 1. The signal line is not plotted — only the slot's RSI line is — but the signal drives the slot's color and its per-slot alerts. See Visuals and Logic for the color rule.

Cost in both directions. Longer signals make the slot's color flip more slowly and the slot's alerts fire later after a regime change. Shorter signals flip more often and can chatter. At signal length 1, many MA families place the signal very close to, or equal to, the slot RSI. Equality does not create a bullish or bearish alert. The plot also falls to its faded/down tone because the color rule is strict on RSI > Signal when a signal value exists.

Line Width

Per-slot plot line width. Default 2, minimum 1.

Blended Weight

Contribution of this slot to the weighted-mean blend. Defaults: 33.3 for slots 01–03, 0.0 for slots 04–05. Minimum 0.

What weight does and does not do. Weight participates in the blend's weighted mean. A weight of zero removes the slot from the blend — but does not turn the slot off. The slot still computes (when enabled), still plots (subject to Hide Plot), and its per-slot alerts still fire.

Cost in both directions. Heavier weights pull the blend toward the slot more aggressively; lighter weights make the slot a minority voice. A five-slot configuration with equal weights is a deliberate choice; a five-slot configuration with arbitrary weights is sometimes an accidental one.

Per-slot Power User block — earned access

Every slot exposes a Power User block for both the RSI smoothing pass and the signal pass. Each block contains:

• On Bar Close? (shared between the RSI pass and signal pass for that slot). Full treatment in MTF and Repainting. Per-slot, not global — this is the STR-specific distinction from Base.

• ALMA parameters: Floor Offset, Offset, Sigma.

• KAMA / FRAMA parameters: Fast, Slow.

• Jurik parameters: Phase, Power.

• Laguerre: Alpha.

• VAMA: Volume Length.

Each family has its own defaults inside the Axiom MA library. The Power User block is exposed for readers who have a specific hypothesis about how a family should behave and want to tune toward that hypothesis.

Philosophy. The Power User block is not more tool; it is narrower tool. Touching it without a stated expectation ("I expect Jurik Power of 2 instead of 1 to make the smoothing more responsive to short bursts") usually produces a configuration that is different, not better. The earned-access framing is not gatekeeping — it is a working description of how much knowledge these parameters require to improve a configuration instead of subtly degrading it. ALMA sigma, Jurik phase, and FRAMA's fast/slow bounds are the kind of parameters where a meaningful tuning hypothesis takes weeks to form and minutes to bungle.

The healthy use pattern is this. Start with defaults. Read the pane at defaults long enough to feel what the smoothing character is doing to your slots. Identify one specific behavior you wish was different and can name out loud — "the Jurik-smoothed slot is slower to flip than I want during hard reversals." Then open the block, change the one parameter that shapes that behavior, and observe whether the pane reports the change you predicted. If it did, you have learned one thing about one parameter. If it did not, you have learned something more important.

The Axiom MA library manual is where the family internals and their tuning are taught; this pack is deliberately not a second copy of that teaching. If you are tempted to tune a Power User block without being able to say what you expect, leave it closed. The tool does not owe you feature-density.

Cross-reference for all MA families and their Power User parameters: see the Axiom MA library manual (Pro trim).

Oscillator reference levels

Two inputs govern the reference lines:

• Overbought Level. Default 70, step 0.1.

• Oversold Level. Default 30, step 0.1.

Both are horizontal dashed lines on the pane. They are visual bumpers. Nothing about the alert list, the coloring, the structure features, or the divergence engine reads these values as thresholds. There is no "blend crossed 70" alert condition. There is no "slot 03 entered oversold" alert. The 30 and 70 lines are there to help you orient on the bounded axis; they are not triggers. Moving them does not change what the tool fires on.

If you were hoping for threshold-crossing alerts wired to these levels, they are not present in the current script. See Alerts for the full inventory.

Blend core

• Plot Blend. On by default. Toggles visibility for both the blended RSI line and the blended signal line. The filled area between the two draws when the blend is plotted.

• Line Width. Default 3. Applies to both blend plots.

The blend color rule (lime when blend RSI is above blend signal, red otherwise) is taught in Visuals and Logic. The code includes a defensive 50-line fallback only if the blend signal is truly unavailable; ordinary reads are blend RSI versus blend signal.

Blend smoothing (master smoothing)

• Enable Smoothing. Off by default.

• Smooth Type. MA family for the master pass. Default EMA.

• Smooth Len. Default 3, minimum 1.

• Power User block. ALMA / KAMA / FRAMA / Jurik / Laguerre / VAMA parameters, same structure as the per-slot block.

When enabled, master smoothing applies a single MA pass to both the blended RSI and the blended signal after the weighted-mean aggregation, then re-clamps into 0..100.

Cost in both directions. Master smoothing produces a calmer, slower blend. The two blend-state alerts fire later, not earlier. Structure features that read the blend — divergence, Keltner, BBWP, and Donchian — inherit the smoothed blend when this is enabled. Per-slot alerts and all-slot alignment alerts read the per-slot states upstream and are unaffected by master smoothing.

Misread to avoid. Calmer is not truer. If the pane was choppy, the chop often carried information — disagreement between slots that the weighted mean was faithfully passing through. A reader who enables master smoothing because the blend "looked noisy" and then reads the smoothed alerts as more reliable is paying a timing cost to remove information they might have needed. The honest use case for master smoothing is a reader who trades on a cadence slower than the blend's natural reactivity, where intra-session chop is genuinely noise for that reader's workflow. That is a real use case. "The pane would look nicer" is not.

Cross-reference: Axiom MA library manual (Pro trim) for the family-specific Power User parameters.

Blend structure features

Four features derive from the blended RSI line. Each has its own enable toggle, its own sizing inputs, its own Power User block, and its own named misuse.

Read order on this page matches the deliberate read order on the pane: Donchian, Keltner, BBWP, divergence. If you are configuring all four, configure them in this order — the ones later in the list are progressively more indirect reads of the blend.

One framing before the four blocks. Every feature below wraps or reads the same blended line. None of them wraps price. None of them is independent of the others. You are not configuring four detectors; you are configuring four framings of one line. The value of running all four is that each framing answers a different question about that line — range pressure, stretch, width regime, pivot geometry. The risk of running all four is that their co-movement looks like confluence and invites the reader to weight their agreement as evidence. The risk is manageable once it is named; the settings below name it once per feature so the framing is installed as you configure.

Blended Donchian

• Show DC. On by default.

• DC Len. Default 20. Lookback for the blend's highest and lowest.

• Basis Type. MA family for the midpoint smoothing. Default SMA.

• Basis Len. Default 1. Length 1 means the basis is the raw midpoint between upper and lower; longer lengths smooth it.

• Show Basis. Off by default.

• Show Fill. On by default.

• Line Width. Default 2.

• Fill transparency. Fixed in the source; there is no separate input.

• Power User block. For the basis MA family parameters.

What it reports. The blended RSI's own highest and lowest over the channel length, stepped. When the blend prints a fresh extreme, the stepline moves to it. A "Donchian press" is the blend sitting against its own recent range edge. That is a description of the blend's behavior against itself — not a price breakout, not an exhaustion call.

Named misuse. Reading a Donchian touch as a price breakout signal. The Donchian here is wrapped around the blend, not price.

Cross-reference: Axiom MA library manual (Pro trim) for the basis MA family Power User parameters.

Blended Keltner

• Show Keltner. Off by default.

• KC Length. Default 20. Lookback for the range EMA on the blend's bar-to-bar absolute change.

• KC Mult. Default 2.0. Multiplier on the range to produce the offset.

• Basis Type. Default EMA.

• Basis Len. Default 20.

• Show Basis. Off by default.

• Show Fill. On by default.

• Line Width. Default 2.

• Fill transparency. Fixed in the source; there is no separate input.

• Power User block. For the basis MA family parameters.

What it reports. A basis (a smoothed blend line) with bands above and below it, spaced by the range EMA times the multiplier. The envelope describes how stretched the blended RSI is against its own smoothed basis, measured by its own recent bar-to-bar range.

Named misuse. Reading a Keltner upper-band touch as "the blend is overbought" or "a reversal is coming." A touch means stretch against the blend's own history. It does not mean exhaustion, and it does not transfer into price-space semantics.

Cross-reference: Axiom MA library manual (Pro trim) for the basis MA family Power User parameters.

Blended BBWP

• Show BBWP. On by default.

• Length. Default 20. Used for both the Bollinger basis and the standard-deviation window on the blend (matched-length Bollinger construction).

• Lookback. Default 252. Number of prior blend-derived widths used to rank the current width. The percentile-rank helper returns nothing until the full window is populated.

• Threshold. Default 50.0. Color split between below-threshold (blue) and above-threshold (aqua) columns.

• Basis Type. Default SMA. MA family for the Bollinger basis; classical Bollinger uses SMA.

• Power User block. For the basis MA family parameters.

What it reports. A percentile rank, on a 0..100 scale, of the blended RSI's current Bollinger-band width against its own recent history. A tall column means the blend's width is ranking high against its own history. A short column means it is ranking low.

Named misuse — the most common on this pane. BBWP is not a price volatility read. The whole measurement is scoped to the blend. A quiet-looking price session can produce a tall BBWP column if the blend itself has been widening; a loud-looking price session can produce a short BBWP column if the blend has been narrow by its own standards. The qualifier "from the blend, not from price" should ride every mention.

Cross-reference: Axiom MA library manual (Pro trim) for the basis MA family Power User parameters.

Blended Divergence

• Show Div. On by default.

• Pivot Len. Default 20, minimum 1. Used as both the left and right pivot strength.

• Plot On Pivot? Off by default. Off places the triangle at the right-shoulder confirmation bar. On back-shifts the triangle by Pivot Len bars to the original pivot.

What it reports. A confirmed-geometry flag. The engine waits for a price pivot to confirm (Pivot Len bars of price action on each side of the pivot), pairs the confirmed pivot with the blended RSI value at the corresponding offset, and compares to the prior confirmed pivot pair in the same direction. Bullish: lower-low in price, higher-low in blend. Bearish: higher-high in price, lower-high in blend. Both evaluated only on confirmed bars.

Named misuse, and the most important on this page. A confirmed divergence triangle is geometry — two price pivots paired with two blended-RSI values at matching offsets. It is not an entry. It is not a turn call. It is not a forecast. Acting on every triangle is a cost this pack actively argues against. The triangle is a prompt to inspect the underlying slot stack and the two pivots that produced the geometry; it is not a replacement for that inspection.

Pivot Len is not a dial for "see more divergences." Lower values produce more triangles because less price action is required on each side of a pivot to confirm it. The result is triangle spam — more markers, none of them stricter. Readers who drop Pivot Len from 20 to 3 or 5 and treat the resulting marker density as new information are paying for volume by losing strictness.

Plot On Pivot honesty beat. The setting is a visual choice, not a timing promise. Plot On Pivot ON back-shifts the marker to the original pivot bar. The alert still fires on the confirmation bar, and you did not have that marker in real time at the back-shifted bar. ON is honest about where the geometry formed; ON is misleading-looking about when you could have acted. Pair the two truths any time you write or teach about this setting.

Where to go next

When you are ready to read the pane rather than configure it, Visuals and Logic installs the four-stage reading order and teaches the co-movement truth about the four structure features. If you are about to touch any bar-close setting, open MTF and Repainting first. If you want to wire an alert, Alerts names exactly what fires and what does not.