Settings

This page is a working reference for every high-impact control in Axiom CVD Osc STR. It is organized by the job the control does, not by the order the Inputs dialog happens to render it. That ordering matters, because the dialog renders five slot groups, five Power User groups, a general oscillator block, the Blend Core block, four optional-feature blocks (Divergence, Keltner, BBWP, Donchian), and four more Power User blocks — and if you try to configure the indicator by walking the dialog top-to-bottom, you will change too many things before you understand which one actually did what.

A useful configuration session looks like this: decide what the slots are for, set the slot identity and timeframes, decide on window mode, tune estimation character only if defaults fail your reading, set weights, touch the blend only if the blend needs a master pass, then handle each structure feature one at a time with the feature's own default view up before you change anything. This page follows that order.

Every field below is named by its user-facing label. For each one, you will find what it does, the tradeoff, what it changes downstream, and a misuse warning where one is earned.

Shape of a reading session with STR

Before the field reference, a small map to hold in your head.

• Slot identity: Enable, Hide Plot, Line Width, Optional Ticker. Purely about what this slot is and whether it draws.

• Windowing: TimeFrame, Lower TF Precision, Window Mode, Window Length, Window Boundary. About when the slot accumulates, how long its accumulation window is, and how the reset marker is drawn.

• Estimation character: Pressure Sensitivity, Wick Weight. How decisively the participation model commits. Defaults are calibrated. Change these last, not first.

• Smoothing: CVD Length, CVD Type, Signal Length, Signal Type, plus Power User fields per MA family. Both per-slot and master smoothing on the blend.

• Blend role: Blended Weight. How much the slot steers the blend.

• Oscillator block: Overbought Level, Oversold Level. Guide lines only.

• Blend Core: Plot Blend, Line Width, Enable Smoothing, Smooth Type, Smooth Len.

• Structure features: Divergence, Keltner, BBWP, Donchian — each with its own teaching load and its own Power User block.

Per-slot controls (CVD 01 through CVD 05)

The five slots are identical in control structure. The defaults are the only thing that differs between them.

Enable CVD 0N

Turns the slot on or off. A disabled slot does nothing — no contribution to the blend, no per-slot alert activity, no drawing. Defaults: slots 1–3 enabled, slots 4–5 disabled.

Misuse warning: enabling all five slots with default weights is a quiet way to arrive at a configuration you did not intend. The blend becomes a weighted mean of five contexts you may not have chosen with care. If slot 4 or slot 5 has no specific question assigned to it, leave it off.

Hide CVD 0N Plot

Hides the slot's own line without disabling the slot. The slot's math still runs, it still contributes to the blend with its assigned weight, and it still fires its own per-slot alerts. Default: ON for slots 01..03 and OFF for slots 04..05; the first-load pane is still clean because slots 04 and 05 are disabled by default.

Misuse warning: this is the most common source of "my blend is reacting to something and I cannot see what." The slot is hidden; it is not gone. If you want to genuinely remove a slot's influence, either disable the slot or set its weight to 0.0.

TimeFrame

The higher timeframe the slot runs on. Must be greater than or equal to the chart timeframe. The indicator throws a runtime error if you set a slot lower than the chart — that is a guardrail, not a bug, and it is there so you cannot ask for higher resolution than the chart bar can carry. The default 5/15/60 stack reads like chart cadence, one step up, and one regime up when you start from a 5-minute chart; on other chart timeframes, check the relationship explicitly.

Defaults: slot 1 = 5 (5-minute), slot 2 = 15 (15-minute), slot 3 = 60 (60-minute), slots 4 and 5 are blank (which falls back to the chart timeframe when enabled). The three enabled slots together give you chart cadence, one step up, and one regime up.

Tradeoff: higher timeframes carry more smoothing and more context, but also more lag in when pressure shifts show up. Lower timeframes carry less lag and more noise. Stacked, the slots let both readings exist on the same pane without choosing one over the other.

Lower TF Precision

The lower timeframe used inside each slot bar to estimate signed volume from intrabar OHLCV. Must be strictly lower than the slot's own timeframe. If it is not, the indicator throws a runtime error with the slot label.

If intrabar data is available for the requested precision, each slot bar's delta is summed from the lower-TF intrabars. If intrabar data is missing — which happens on older history after TradingView stops serving intrabars at that precision — the slot falls back to estimating delta from the slot bar's own OHLCV. The fallback is a cruder read. The line does not visually announce the change. The verification move here is to scroll back past the intrabar history window and look for the point where the slot line behaves more abruptly; that is the fallback speaking.

Misuse warning: pushing precision very low on a slot you then backtest across a long history will give you a slot that reads at its requested precision on recent bars and at a cruder precision on older ones. If your workflow cares about the consistency of the estimator across history, pick a precision that the slot's whole chart can actually carry.

Window Mode

• Session resets accumulation at each window anchor. The reset is visible as a dashed vertical line in the slot's color (if Hide Plot is OFF). This is the mode to use when you think of delta as a per-session accounting.

• Rolling keeps a sliding lookback the size of the window. There is no reset marker — by design, because the window is a sliding window rather than a reset event. This is the mode to use when you think of delta as a continuous pressure read across time.

Both are legitimate. They produce different lines from the same data. There is no "best" choice; there is only the choice that matches how you want to read.

Window Length

The session anchor in Session mode (typically D for daily, W for weekly, or a session timezone string), or the lookback duration in Rolling mode. Must be at least as large as the slot timeframe. The indicator throws a runtime error if it is not.

Default: D. On an intraday chart in Session mode, this produces a fresh anchor at each day's session boundary for the instrument's exchange.

Window Boundary

Color for the dashed Session reset marker on a visible slot. It is a drawing control only. It does not move the reset, change the accumulation window, or affect the CVD read.

CVD Length and CVD Type

Smoothing length and MA family applied to the normalized CVD line. The normalization produces a 0-to-100 value against the window's observed high and low; this smoother is what actually draws as the per-slot CVD line.

Default: length 3, type SMA. Short and fast.

Some MA types open additional Power User parameters farther down in the dialog:

• ALMA uses CVD ALMA Floor Offset?, CVD ALMA Offset, CVD ALMA Sigma.

• KAMA and FRAMA use CVD KAMA/FRAMA Fast and CVD KAMA/FRAMA Slow.

• Jurik uses CVD Jurik Phase and CVD Jurik Power.

• Laguerre uses CVD Laguerre Alpha.

• VAMA uses CVD VAMA Vol Length.

For what each MA family actually does and what the Power User parameters change, link out to the Axiom moving-average library manual. This page does not reteach the MAs — they have their own reference, and duplicating it here would drift as the library evolves.

Misuse warning: there is no "best" MA type for the asset class you trade. Different MAs have different character. Pushed to extreme parameters, the non-SMA families can behave like crossover machines on every minor swing. That behavior belongs to the MA, not to the delta.

Signal Length and Signal Type

Length and MA family for the signal line. Same conventions as the CVD line, smoothed on top of the already-smoothed CVD. Default: length 3, type SMA.

The signal is what the slot line crosses to produce its per-slot bullish or bearish state. If you want the state less twitchy, raise the signal length before you raise the CVD length.

Line Width

Thickness of this slot's plot. Default 2.

Blended Weight

The slot's share of the weighted average that produces the blend. Normalized across the enabled slots at runtime — so you can express weights in any unit you like; the blend uses the ratio.

Defaults: slots 1–3 are 33.3 each, slots 4–5 are 0.0. A slot with weight 0.0 is still active, still fires its own per-slot alerts, and still draws if Hide Plot is OFF. It simply does not steer the blend.

Use case for weight 0.0: you want the slot's alert surface without letting it pull the blend. For example, a correlated-instrument slot that you want to watch but do not want merged into your core read.

Misuse warning: equal weights across five mismatched slots is a quiet way to flatten the blend. If one slot is on a much faster cadence than the others, it will fight the slower ones at most moments and the blend will sit near 50 more often than it should.

Pressure Sensitivity

A 0.25 to 4.0 tunable, step 0.05, default 1.50. This is one of two controls that shape the character of the participation model that runs underneath every slot.

What it changes: how decisively the model commits participation on each bar. Higher values make the line more decisive — faster to commit, more staccato, more willing to treat a bar as a clear participation event. Lower values make the line more forgiving — slower to commit, smoother across indecisive stretches, more willing to carry prior state forward across bars whose intent is not obvious.

What it does not change: the inputs the model reads (body direction, close location, wick asymmetry). Those are always in play. The control adjusts how decisively the inputs combine into a classification, not which inputs exist.

Verification: push Pressure Sensitivity to 4.0 on one slot and watch the line behave more abruptly on indecisive bars — short bodies and balanced wicks that previously faded through partial carry can now commit more readily. Drop to 0.25 and watch the same stretch produce a calmer line; more bars fall into the indecisive carry path, and the carried state is softer. The difference should be obvious inside a single chart session on a liquid instrument. If it is not, you have changed the value on a hidden slot or a slot whose weight is doing nothing in the blend.

Misuse warning: there is no "best" Pressure Sensitivity for the asset class you trade. Defaults are calibrated against a broad mix. Move in small steps (0.10 or 0.20 at a time) and give a change a real session before changing again. The temptation when a backtest disappoints is to push the dial harder; the cost of pushing it harder is that the slot becomes more sensitive to bars whose participation was genuinely ambiguous, and you trade smoother backtest output for noisier real-time behavior. Do not expect this page to publish the internal mapping from the knob value to the classifier thresholds — that boundary exists to keep the model useful rather than cloneable. For the Geeks explains the shape of the mapping at mental-model altitude.

Wick Weight

A 0.0 to 0.50 tunable, step 0.05, default 0.20. The second control shaping the participation model.

What it changes: how much wick asymmetry enters each bar's intent score. At 0.0, wicks are ignored and the bar's body and close location carry the whole read. At 0.50, wicks are given substantial influence — a wick-heavy rejection will move the slot more than a similar-looking close-to-close move on a bar with tight wicks.

What it does not change: the presence or absence of body and close-location reads. Those stay in play at every setting.

Verification: compare a chart stretch with several wick-heavy rejection bars at Wick Weight = 0.0 and again at Wick Weight = 0.50. The line through the same bars will look different.

Misuse warning: same pattern as Pressure Sensitivity — small steps, full sessions, and no expectation that the page will hand you the body and close-location coefficients.

Optional Ticker

If set, the slot runs against that symbol instead of the chart symbol. Leave blank to run on the chart symbol. Dependent on dynamic_requests = true in the indicator declaration, which is on.

Misuse warning: a slot pointed at a different exchange's session will draw its session reset marker on that exchange's boundary, not the chart's. That is expected behavior, not a bug, and it means a cross-exchange slot needs you to hold the session mismatch in mind while you read it.

Per-slot Power User (CVD 0N PU)

On Bar Close?

The repaint switch for the slot. Default ON.

• ON reads the prior slot-bar value via a one-bar shift. Slower but settled. The slot's output does not change inside the current slot bar.

• OFF reads the live slot-bar value. The slot's output updates in real time and can change until the slot bar closes.

Both are legitimate. ON is the honest posture for backtesting and for historical review. OFF is useful when you want the live read and you understand the cost. The MTF and Repainting page walks the tradeoff at full depth; that is where you should sit with this switch before you move it.

CVD and Signal Power User fields

• CVD ALMA Floor Offset?, CVD ALMA Offset, CVD ALMA Sigma: active only when CVD Type is ALMA.

• CVD KAMA/FRAMA Fast, CVD KAMA/FRAMA Slow: active only when CVD Type is KAMA or FRAMA.

• CVD Jurik Phase, CVD Jurik Power: active only when CVD Type is Jurik.

• CVD Laguerre Alpha: active only when CVD Type is Laguerre.

• CVD VAMA Vol Length: active only when CVD Type is VAMA.

The same block repeats with a Signal prefix for the signal-line smoother. Fields are inert when the corresponding MA type is not selected; the dialog does not grey them out, but they do nothing until the MA type is set.

For what each field changes, cross-reference the Axiom moving-average library manual.

Oscillator group (general)

Overbought Level

Default 80. Upper dashed guide line in the pane. Purely visual. No alert is generated at this level. Do not read a crossing as a signal.

Oversold Level

Default 20. Lower dashed guide line. Same posture as Overbought — visual reference only.

These guides are worth keeping at meaningful defaults because the pane's boundedness means every slot uses the same scale. If you move them, move both, and move them for the same reason.

Blend Core

Plot Blend

Draws the blended CVD line, blended signal line, and the fill between them. Default ON. If you are studying only per-slot behavior for a session, turn this off to reduce pane clutter.

Line Width

Thickness of the blend plots. Default 3. Thicker than slot lines by default because the blend is the pane's primary summary read.

Enable Smoothing

Adds one more MA pass on top of the blend line and the blend signal. Default OFF.

Tradeoff: master smoothing reduces short-term noise at the cost of lag. If the blend is twitching and the twitch is obscuring the read you care about, this is the right control. If the blend is twitching because the slots themselves are fighting, this will hide the fight without resolving it — and the better move is usually to reconsider which slots are in the stack.

Misuse warning: turning Master Smoothing on early, while you are still learning what the slot stack is doing, will mask behavior you would benefit from seeing. Leave it off until the blend is doing something specific you want to calm down.

Smooth Type

MA family used for the master pass. Default EMA. Same families as the per-slot smoothing.

Smooth Len

Length of the master pass. Default 3. Short by default — the master pass is meant to calm the blend, not to reshape it.

Blend Div (the divergence module)

STR's highest-stakes feature. The one whose misreadings hurt the most, and the one most likely to attract attention from a reader who only meant to glance at the pane. Read this subsection carefully and read it twice if anything in it surprises you on the first pass.

Show Div

Turns the divergence module on or off. Default ON.

When on: the module watches for confirmed price pivots and compares them to the blended-CVD value at the same offset. A confirmed bullish divergence is two confirmed price pivot lows in lower-low order paired with two blended-CVD values in higher-low order at the same offsets. Bearish is the mirror. When the condition is met inside a confirmed bar, the module marks a triangle and fires the corresponding alert.

Pivot Len

Default 20. Pivot strength used as both the left and right strength for the pivot functions on chart price and for reading the blended-CVD value at the offset.

What it changes: higher values require more bars to the left and to the right of the pivot before the pivot is confirmed. That means fewer triangles, stricter geometry, and longer confirmation delay. Lower values require fewer bars, produce more triangles, and confirm faster.

Misuse warning: lowering Pivot Len to "see more divergences" does not produce better divergences. It produces more triangles at a noisier scale. A triangle at Pivot Len = 5 and a triangle at Pivot Len = 30 are both geometrically valid descriptions of their own pivot definitions, but they are not the same event and the lower setting will print triangles through regions where the structure is merely noisy. There is no "best" pivot lookback — the pack will not name one. Choose a value that matches the swing scale you actually read.

Plot On Pivot?

Default OFF. The single most important honesty control on the indicator. Treat it as a control whose two settings are not interchangeable views of the same data — they are two different visual contracts with the reader.

• OFF: the triangle prints on the confirmation bar — that is, the bar where the right-hand pivot completes. The triangle visually sits Pivot Len bars after the actual pivot. This is the honest real-time posture. The script could not have flagged the triangle before the pivot confirmed, and the marker appears on the bar where it became true.

• ON: the triangle is back-shifted by Pivot Len bars to sit on the original pivot bar. The geometry is easier to study because the marker visually lands where the pivot actually occurred in the chart. The alert still fires only on the confirmation bar. The marker was not visible at the back-shifted bar in real time. The visual is correct about where the geometry formed and silent about when the script could have known.

This is the cleanest example on the indicator of a control whose behavior differs between live reading and historical study. Leave OFF as your working default. Use ON only for post-session study, and only after you have internalized what the marker's position means in each mode.

The teaching cost of getting this wrong is not abstract. A reader who learns the pane with ON builds an instinct that the triangle appears at the bar where it visually sits. The instinct feels right because the geometry is right. The instinct fails on the first live trade, because in live trading the triangle does not appear at the back-shifted bar — there is nothing visible there until Pivot Len bars later. If you teach someone else to read the pane, leave OFF during the teaching session. The polish pass on this pack flagged this as a recurring drift point; the warning is repeated on multiple pages on purpose.

Blend KC (Keltner envelope)

Show Keltner

Turns the envelope on or off. Default OFF.

When on: upper and lower envelope lines wrap the blended CVD line. An optional basis line can draw between them. The envelope is derived from the blend, not from price.

KC Length

Default 20. Lookback used to smooth the one-bar absolute change of the blended CVD line for the envelope's range component.

KC Mult

Default 2.0. Multiplier on the smoothed range for the upper and lower envelope distance. Higher values give a wider envelope. Lower values give a tighter one.

Misuse warning: do not move KC Mult until you have lived with the default for at least a session. The default is deliberately a familiar Keltner width.

Basis Type

MA family used for the envelope's basis — the smoothed midpoint of the blend itself. Default EMA.

Basis Len

Length of the basis filter. Default 20.

Show Basis

Draws the basis line. Default OFF. Turn on if you want to see the midpoint the envelope is computed around.

Show Fill

Fills between the upper and lower envelope lines. Default ON. The fill uses a fixed subtle transparency in the source; there is no separate fill-transparency input.

Line Width

Thickness of the upper and lower envelope plots. Default 2.

The Keltner envelope is a stretch envelope around the blend. A touch of the upper line says the blend has stretched far from its own basis by its own recent range. A touch is not exhaustion, not overbought, not an exit. The blend can ride the upper line for long stretches under sustained pressure. Read the envelope as distance, not as regime.

Blend BBWP (Bollinger Band Width Percentile)

Read the qualifier once and hold onto it: BBWP on this indicator is measured on the blended CVD line itself, not on price. It is the blend's own band-width percentile, ranked across the blend's own recent width history. Every sentence below operates inside that qualifier.

Show BBWP

Turns the BBWP columns on or off. Default ON.

Length

Default 20. Used for both the Bollinger basis and the standard deviation on the blended CVD line. Keeping these matched is the standard Bollinger construction — the pack does not separate them.

Lookback

Default 252. Number of prior BB-width values ranked to compute the percentile. Longer lookbacks produce a stabler regime read; shorter lookbacks produce a more reactive one.

Threshold

Default 50.0. Percentile level where the column color switches between the below-threshold color and the at-or-above-threshold color.

Misuse warning: the threshold is a display control. It does not change what the percentile is measuring; it changes where the visual hue flips. Move it if you want the visual to emphasize different parts of the percentile range. Do not move it thinking you are changing what the columns "mean."

Basis Type

MA family used for the Bollinger basis on the blended CVD line. Default SMA. Classical Bollinger uses SMA; STR matches that default.

Blend DC (Donchian channel)

Show DC

Turns the Donchian channel on or off. Default ON.

DC Len

Default 20. Lookback used to find the highest and the lowest of the blended CVD line for the upper and lower stepline channels.

Tradeoff: longer lookbacks mean the channel changes less often and the "new extreme" events are genuinely rarer. Shorter lookbacks mean more frequent channel resets and more frequent "new extreme" touches, most of which are minor by any longer-horizon read.

Basis Type

MA family used to smooth the midpoint when basis is shown. Default SMA.

Basis Len

Length of the basis filter. Default 1. At 1, the basis is effectively unsmoothed.

Show Basis

Draws the smoothed midpoint as a stepline. Default OFF.

Show Fill

Fills between the upper and lower stepline channels. Default ON. The fill uses a fixed subtle transparency in the source; there is no separate fill-transparency input.

Line Width

Thickness of the upper and lower stepline channels. Default 2.

The Donchian channel marks where the blend has been pushing against its own highs and lows. A touch of the upper stepline says the blend has just printed a new local high over the channel length. That says something about the blend's own recent range. It does not say anything about where price will go next.

Blend Smooth PU, Blend KC PU, Blend BBWP PU, Blend DC PU

Each of these groups exposes the standard MA-family-specific Power User controls (ALMA Floor / Offset / Sigma, KAMA/FRAMA Fast / Slow, Jurik Phase / Power, Laguerre Alpha, VAMA Vol Length). They are only meaningful when the corresponding feature's MA Type selection requires them. The dialog does not grey them out; they sit inert until their MA family is selected.

For the behavior of each Power User field, cross-reference the Axiom moving-average library manual rather than reteaching the families here. The cross-reference is deliberate: if the MA library changes, that manual is where the changes are reflected. Duplicating the MA teaching in this page would decay.

Settings cheat-sheet, grouped by the job they do

Where to go next

• For how the settings above read on the chart, and for walkthroughs of the ambiguous states they create, Visuals and Logic.

• For the timing and repaint tradeoff that lives under On Bar Close? — and the confirmation timing that lives under Pivot Len and Plot On Pivot? — MTF and Repainting.

• For the estimate boundary and the structure-features-are-not-independent fact that matters when you look at more than one feature at once, Limitations and Trust Boundaries.

• For the participation model's shape and each structure feature's construction at mental-model altitude, For the Geeks.