Std Error Bands Indicator: Formula, Settings and How to Read It
Standard Error Bands (Jon Andersen, 1996) surround a linear regression line with bands set at a multiple of the standard error of the estimate, then smooth all three lines. Senzoukria uses Andersen's defaults: a 20-bar regression, a multiplier of 2 and a 3-period smoothing.
Senzoukria · Indicators · Updated September 2026
Std Error Bands ships with the Senzoukria desktop app, in the Averages & volatility group of the indicator catalogue. It is drawn on the price chart.
What Std Error Bands measures
A least-squares line is fitted to the closes of the window and the basis is its terminal value at the current bar. The band distance is the standard error of the estimate: the root mean square of the residuals around that fitted line, divided by N minus 2 rather than by N or N minus 1, because two parameters, the intercept and the slope, were estimated from the data. What the bands measure is therefore how tightly price has followed its own trend line over the window, not how far it has travelled. The period is constrained to at least 3, since at N equal to 2 the line passes exactly through both points and the divisor would be zero. When the residuals are genuinely zero with N of 3 or more, the three lines coincide, which is a true reading rather than a gap.
The formula, as implemented
This is not a description of how the indicator is usually defined elsewhere — it is what the shipped code computes, documented next to the implementation:
Standard Error Bands (Jon Andersen, 1996) : fit y = a + b·t (moindres carrés) sur la fenêtre N ; basis = a + b·(N−1) (point terminal de la droite) SEE = √( Σ (y_t − (a + b·t))² / (N − 2) ) ← erreur-TYPE de l'estimation : N−2 degrés de liberté (2 paramètres estimés, a et b), PAS ÷N ni ÷(N−1) upper/lower = basis ± k·SEE, puis les TROIS lignes lissées par SMA_s. POURQUOI CE N'EST PAS `linreg-channel` (averages2.ts) : celui-ci utilise l'écart-type de POPULATION des résidus (÷N, convention Bollinger) et ne lisse rien. Ici c'est l'erreur-type au sens statistique (÷(N−2)) suivie du lissage 3 périodes d'Andersen : bandes plus larges, contour beaucoup plus calme. Deux lectures différentes du même fit, deux indicateurs. N ≥ 3 imposé (N−2 ≥ 1 : à N = 2 la droite passe exactement par les deux points, Σrésidus² = 0 ET N−2 = 0 — un 0/0 sans réponse honnête). Résidus nuls avec N ≥ 3 (données parfaitement alignées) → SEE = 0 : les trois lignes se confondent, ce qui est la vérité, pas un trou. Warm-up (fenêtre incomplète, puis lissage) → trou synchronisé. Défauts N=20, k=2, lissage=3 (Andersen).
How to read it
- Narrow bands mean the closes have clung to the fitted trend line; wide bands mean they have scattered around it, whatever the direction of that line.
- Watch the basis and the width separately: the basis reports the direction of the fit, the width reports how well the fit describes the window.
- A band that tightens while the basis keeps its slope describes a move that has become more orderly, which is a description of the past window and not of the next one.
- All three lines are smoothed over 3 periods, so their contour is calmer than the raw fit and a change in width shows up a couple of bars late.
Parameters and defaults
Period defaults to 20 with a minimum of 3 and a maximum of 500, the multiplier to 2 in a 0.5 to 5 range, and the smoothing to 3, Andersen's published value, adjustable up to 50. Raising the smoothing produces a visibly quieter envelope at the cost of further delay; setting it to 1 disables smoothing altogether and exposes the raw standard error, which is considerably more jagged.
| Parameter | Type | Default | Range |
|---|---|---|---|
| Period | number | 20 | 3 – 500 |
| SE multiplier | number | 2 | 0.5 – 5 |
| Smoothing | number | 3 | 1 – 50 |
What it does not show
These bands are not Bollinger Bands in disguise: they measure dispersion around a sloping regression line, so a strong steady trend can keep them narrow while price travels a long way, a case where a standard-deviation band would be wide. The separate linear regression channel in the catalog fits the same line but uses a population standard deviation of the residuals and applies no smoothing, which makes it a different indicator with a different width. Period and smoothing both extend the warm-up, and the envelope is drawn only where all three lines exist, so a long period combined with heavy smoothing leaves a wide blank at the left edge of the chart.
Using it in Senzoukria
Add it from the Indicators panel of any footprint chart or candle chart. It runs on futures data from Rithmic or Databento and on crypto pairs from Binance and Bybit, on the same engine — the calculation does not change with the venue, only the data feeding it does. Market data subscriptions are billed by the provider, separately from the app.
Related indicators
- WMA — Averages & volatility
- Hull MA — Averages & volatility
- ATR (Wilder) — Averages & volatility
- Bollinger — Averages & volatility
- Keltner — Averages & volatility
- Donchian — Averages & volatility
See the full indicator library, or start with the order flow guide if you are new to reading aggression, delta and absorption.
Frequently asked questions
- What do Standard Error Bands measure?
- They measure how closely price has tracked its own linear regression line over the window. The basis is the endpoint of a least-squares fit over 20 bars by default, and the bands sit at a multiple of the standard error of that estimate. Narrow bands indicate a well-fitted trend, wide bands a scattered one.
- Why is the standard error divided by N minus 2?
- Because the regression estimates two parameters from the data, an intercept and a slope, which leaves N minus 2 degrees of freedom in the residuals. Dividing by N or by N minus 1 would understate the dispersion. This is also why the period has a minimum of 3: at two points the fitted line is exact and the divisor would be zero.
- How do Standard Error Bands differ from Bollinger Bands?
- Bollinger Bands measure the standard deviation of price around a flat moving average, so a steady trend widens them mechanically. Standard Error Bands measure the residuals around a sloping regression line, so a steady trend can leave them narrow. The two answer different questions: distance from the mean, against quality of fit to the trend.