FreeOpen sourcePine Script v6
Separate above-basis and below-basis dispersion into observation frequency and conditional squared-excursion severity. Inspect stabilized asymmetric wings alongside their attribution, latest mass shift and side-normalized location.
Conventional Bollinger Bands place one standard deviation symmetrically around an average. The primary wings here use independently stabilized side statistics. The measurements share one basis-relative framework rather than combining unrelated oscillators or volume filters.
A basis, asymmetric upper/lower wings, split fill, dominant-wing glow, seven-row readout and sparse excursion/re-entry/cross-wing markers. A conventional SMA-centered symmetric reference is available separately and is off by default.
For each window, deviations equal historical observations minus the current selected basis. Positive and negative squared deviations accumulate separately. Each side’s semivariance equals its squared mass divided by all valid observations: occupancy × conditional squared-excursion severity.
Sparse sides are partly shrunk toward pooled dispersion. Square roots, the multiplier and optional RMA smoothing produce distances from the basis. Unshrunk mass, occupancy and severity statistics remain separate.
ATTRIB explains current asymmetry; SHIFT explains its one-bar change. Upper mass can remain dominant while the latest change moves toward the lower side.
Delta semivariance = delta occupancy × average severity + delta severity × average occupancy. The midpoint split allocates interaction equally. Upper-minus-lower change is normalized by average total semivariance across the two windows, not by price.
MASS, frequency and severity are smoothed independently. Their visible values need not preserve the raw algebraic identity. The Data Window residual checks numerical consistency of the raw shift decomposition.
Use a standard time-based chart. Identify both wings, the basis and readout at default settings. Verify Source and Basis type before changing one input at a time.
Open the public Bollinger Bands Asymmetry Attribution page and add it to TradingView. No SG Group account is required.
Overlapping lines, bands, backgrounds or tables can be duplicate drawings of the same kind of information. Never treat visual overlap by itself as stronger evidence or signal agreement. Conventional Bollinger Bands, Keltner channels and VWAP deviation envelopes can duplicate nearby lines and fills. The optional classic reference is also a different calculation. Identify which geometry a marker crossed.
Read MASS, ATTRIB, SHIFT, STRETCH, WINGS and NOW in order to separate level, driver, change, location, geometry motion and realized containment.
Cyan and magenta MASS shares sum to 100%. F is occupancy; S is conditional squared-excursion severity. ▲ favors upper, ▼ lower, · is below threshold, and S? indicates inadequate comparison samples.
SHIFT F↑/S↑ moves the directional mass difference upward; F↓/S↓ moves it downward. These arrows do not prescribe trade direction.
STRETCH divides source-to-basis distance by the same-side wing distance. +1.00x and −1.00x are the upper and lower wings; ±0.50x is halfway. Absolute stretch above one is outside. Default edge starts at 0.8.
WINGS U/D shows upper/lower distance expanding, contracting or stable. The ratio is upper distance divided by lower distance. NOW reports current location or a live transition preview; C is past realized containment.
Execution status on a forming/closed bar. Confirmation does not freeze the live readout.
Neither driver is material / frequency-only bias / severity-only bias.
Frequency and severity reinforce / oppose each other. These describe composition.
Too few observations on a side for severity comparison.
No meaningful basis-relative dispersion / required history unavailable.
Frequency/severity is the dominant change driver; upper/lower describes the mass-difference direction.
Same-direction reinforcement / shared net direction without dominant driver / opposing components / net shift below threshold.
Upper outside / near upper edge / upper half. Lower states mirror these; On basis is at the center.
Realized containment / upper-to-lower distance ratio; neither is a forecast probability.
The labels below match the published script’s English Inputs panel. Defaults reflect the public source at the time of review and may change in a later TradingView update.
On small screens, each input is reformatted as a readable stacked card.
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Source | close / options: open, high, low, close, hl2, hlc3, hlcc4, ohlc4 | Selects the series analyzed throughout the study. The default is close. hl2 averages high/low; hlc3 averages high/low/close; hlcc4 weights close twice; ohlc4 averages OHLC. |
Attribution length | 20 / min 5, max 250 | Sets the rolling window, default 20 observations. Each window compares historical values with its current basis, not each bar’s former basis. |
Basis type | SMA / options: SMA, EMA, RMA, WMA | SMA provides the arithmetic-center interpretation. EMA, RMA and WMA are alternative basis-relative research choices. |
Wing multiplier | 2 / min 0.1, max 10 | Scales the independently stabilized root-mean-square excursions. The default 2 is not a confidence level or a future coverage probability. |
Sparse-side stabilization | 4 / min 0, max 50 | Stabilizes a sparsely populated wing toward pooled dispersion. Zero disables shrinkage when that side has observations. Raw attribution statistics are unchanged. |
Wing smoothing | 2 / min 1, max 20 | RMA smoothing of wing distances; 1 leaves stabilized widths unsmoothed. Smoothing adds lag. |
Show classic symmetric reference | Off | Shows a conventional SMA-centered symmetric reference. It is off by default and uses a different construction from the primary wings. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Attribution smoothing | 3 / min 1, max 30 | Smooths mass, frequency and severity diagnostics independently. The visible smoothed diagnostics do not form the exact raw factorization. |
Dominance entry threshold | 12 / min 1, max 90 | Absolute smoothed mass-bias threshold for entering dominance, default 12. It is not an upper-share level of 12%. |
Dominance release threshold | 7 / min 0, max 89 | Release hysteresis retains an established dominance state through small counter-moves. Keep it below the entry threshold. |
Driver significance threshold | 6 / min 0, max 80 | Absolute bias threshold used to classify frequency and severity drivers. It is a display/classification threshold, not a statistical p-value. |
Minimum observations per side | 3 / min 1, max 25 | Minimum observations on both sides before severity comparison. The effective requirement is capped at half the attribution window. Insufficient samples produce Sparse/S?. |
Containment lookback | 100 / min 10, max 1000 | Lookback for realized containment, default 100. It measures past observations inside displayed wings and does not guarantee future coverage. |
Edge zone starts at | 0.8 / min 0.5, max 0.99 | Defines the edge zone in side-normalized units. Default 0.8 starts at 80% of the matching wing distance; 1.00 is the wing. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Mass shift smoothing | 2 / min 1, max 20 | RMA smoothing for frequency/severity contributions to one-bar directional mass change, separate from level attribution smoothing. |
Mass shift significance threshold | 2 / min 0.1, max 50 | Threshold for materially active normalized shift components. Default 2 is not a 2% price change. |
Mass shift driver dominance ratio | 1.25 / min 1, max 5 | Required component-size ratio for one shift driver to dominate, default 1.25. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Confirmed events only | On | When on, event markers and alerts wait for bar close. Wings and readout remain live. Turning it off permits provisional intrabar events. |
Wing motion threshold % | 0.5 / min 0, max 20 | Minimum one-bar change in wing distance for expansion/contraction arrows, default 0.5%. It does not describe price direction. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Show asymmetric wings | On | Shows the asymmetric upper and lower wing lines. |
Show basis | On | Shows the selected moving basis. |
Show split upper/lower fill | On | Fills the upper and lower regions independently. |
Show state rail | Off | Shows a state-colored rail behind the basis; off by default. |
Show dominant-wing glow | On | Adds a translucent glow to the wing matching the current mass-dominance state. |
Wing line width | 2 / min 1, max 5 | Width of the upper and lower wing lines. |
Basis line width | 2 / min 1, max 5 | Width of the basis line. |
State rail width | 4 / min 2, max 10 | Width of the optional state rail. |
Dominant glow width | 5 / min 2, max 12 | Width of the dominant-wing glow. |
Split fill transparency | 94 / min 0, max 100 | Split-fill transparency; higher values are more transparent. |
State rail transparency | 86 / min 0, max 100 | State-rail transparency; higher values are more transparent. |
Dominant glow transparency | 82 / min 0, max 100 | Dominant-glow transparency; higher values are more transparent. |
Upper wing | rgba(0,229,255,1) | Upper-wing color, cyan by default. |
Lower wing | rgba(255,77,214,1) | Lower-wing color, magenta by default. |
Basis | rgba(255,209,102,1) | Basis color. |
Upper-mass state | rgba(88,255,148,1) | Upper-mass state color. |
Balanced state | rgba(178,120,255,1) | Balanced-state color. |
Lower-mass state | rgba(255,96,121,1) | Lower-mass state color. |
Sparse-sample state | rgba(255,196,74,1) | Sparse-sample state color. |
Classic reference | rgba(153,165,190,1) | Classic symmetric reference color. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Readout mode | Visual / options: Off, Visual | Visual shows the readout; Off hides it. |
Readout position | Top right / options: Top right, Middle right, Bottom right, Top left, Middle left, Bottom left | Positions the readout in one of six left/right locations. Separate it from other panels. |
Readout text size | Small / options: Tiny, Small, Normal | Chooses Tiny, Small or Normal readout text. |
Axis label clearance % | 3 / min 0, max 12 | Transparent clearance between edge-anchored readout and price-scale labels, default 3%. |
Readout background | rgba(15,18,29,1) | Readout background color. |
Readout transparency | 8 / min 0, max 100 | Readout background transparency. Higher values expose more chart content behind the text. |
Readout text | rgba(244,247,255,1) | Readout text color. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Show first outside excursion | On | Marks the first inside-to-outside transition with a triangle, not every outside bar. |
Show first re-entry | On | Marks the first return inside with a circle on the wing. |
Show direct cross-wing jump | On | Shows one basis diamond for a direct jump between opposite outsides, consolidating duplicate excursion/re-entry markers. |
Show dominance-flip markers | Off | Optional basis markers for transitions into upper/lower mass dominance; off by default. |
Create the desired condition in TradingView. Confirmed events only is on by default: markers and alerts wait for close while wings and the readout continue updating.
This tool is for analysis and visualization. It is not investment advice, a trade recommendation, a forecast, or a promise of profit or accuracy. Results depend on inputs, symbol, timeframe, and data feed. You remain responsible for every trading decision.
Primary wings are not conventional Bollinger Bands or confidence intervals. A multiplier of two does not imply 95% coverage. Realized containment is backward-looking. Outside observations alone do not establish reversal, continuation or exhaustion.
Every historical observation in a window is compared with that window’s current basis. EMA/RMA/WMA variants are not conventional semivariance around an arithmetic sample mean.
Insufficient observations on either side produce Sparse/S?. Stabilized smooth geometry does not make the underlying severity comparison adequately sampled.
The study uses no higher-timeframe requests or lookahead, but forming-bar wings and diagnostics change. Disabling confirmation permits provisional markers and alerts. Synthetic charts change the analyzed series.
The Data Window includes bias components, shares, stretch, wing distances, containment, sample adequacy, mass shift and Flow Bridge Identity Residual. Small nonzero residuals can arise from floating-point arithmetic.
MASS, ATTRIB, SHIFT, wings and glow share related statistics. Their agreement is not independent confirmation, a win rate, expected return or future directional probability.
Clarify symmetric versus asymmetric geometry, level versus shift, sparse samples and realized coverage.
No. They use stabilized side-specific conditional excursions. A separate conventional SMA-centered symmetric reference is optional.
No. It is a smoothed share of basis-relative squared dispersion.
No. Sparse-side stabilization changes the drawing, not the raw sample size. Check Sparse and S?.
Yes. MASS is the current level; SHIFT is its latest change. Upper dominance can coexist with a lower-directed change.
It is a free open-source TradingView publication. Review its terms before modification or redistribution.
This guide is grounded in SG Group’s live catalogue, the script’s published TradingView description, and the user-facing inputs in its open source. Labels and defaults can change after a script update.
