Standard error bands
Bands based on the standard error of a linear regression estimate — they narrow when the trend is well-defined and widen when price action is more scattered.
What the numbers mean
Twenty-one is the regression period, 2 sets how many standard errors the bands sit from the fitted line, and the smoothing value averages the result. Standard error measures how tightly bars cluster around the regression, so the bands narrow when a trend is orderly and flare when it is not.
- 21 — the regression period
- 2 standard errors — band distance from the fitted line
- 3 — the smoothing applied to the result
- Standard error measures how tightly bars cluster around the fit, not how far price ranges
How it is used
Band width is the signal: tight bands indicate a well-defined trend worth following, while sudden widening warns that the trend is losing coherence.
- Bands narrowing → the trend is orderly and worth following
- Bands flaring → the trend is losing coherence
- Width, not position, is the signal here
How it relates to everything else
This is the band that inverts the usual reading. With Bollinger Bands, narrow means quiet; here narrow means trending cleanly. Plotting both is a quick way to distinguish a quiet range from a tidy trend — Bollinger narrow with standard error wide is chop, both narrow is a smooth, low-volatility advance.
- Narrow = trending here, the opposite of Bollinger's narrow = quiet
- Built on the linear regression curve, so it slopes with the trend
- Best read as a trend-quality gauge rather than an entry tool
What to watch for
It is easily confused with standard-deviation bands, which behave almost the opposite way — narrow there means quiet, narrow here means trending.