Concept

Bollinger Band Tag Reversion

Bollinger Band Tag Reversion is a Volatility concept. The Library holds 1 implementation — a working definition you can pull into Quant.

Top Bollinger Band Tag Reversion indicator

The top custom implementation, built on the original standard Bollinger Band Tag Reversion formula.

1 total

This Bollinger Band Tag Reversion implementation is strategy-ready: open it in Quant, set your rules, and it backtests automatically.

Bollinger Band Tag Reversion preview

Bollinger Band Tag Reversion

Indicator

What is a Bollinger Band Tag Reversion?

A Bollinger Band tag reversion is a mean-reversion setup: price touches or pierces one of the outer Bollinger Bands (by default two standard deviations from a 20-period average) and the trade bets on a rotation back toward the middle band. The logic is statistical stretch: a tag marks price at the edge of its recent distribution, and in quiet, range-bound conditions that edge often behaves like a soft boundary rather than a doorway.

The bands themselves date to the early 1980s, when John Bollinger replaced the fixed-percentage envelopes then in common use with bands scaled by standard deviation, so band width adapts to volatility instead of staying constant. Fading band tags became one of the tool's two canonical uses, the mirror image of the breakout style that trades a Bollinger Squeeze. In Bollinger on Bollinger Bands (2001) he treats tags as raw material for setups, never as signals in themselves.

That caution is the heart of the concept: a tag of the upper band is not by itself a sell signal, and a tag of the lower band is not by itself a buy signal. In trends, price can walk the band, tagging it repeatedly while the move continues. Reversion traders therefore qualify tags with regime and confirmation: a non-trending backdrop, a close back inside the bands, a rejection candle such as a pin bar, or an oscillator reading like %B turning back from an extreme.

The statistical framing deserves honesty. Two standard deviations of a 20-period sample sounds precise, but returns are not normally distributed and the sample is small, so no fixed probability attaches to a tag. What a tag reliably says is that price is stretched relative to its own recent volatility; whether stretch mean-reverts is a regime question, which is why the setup's edge lives in the filter more than in the touch.

How to identify a Bollinger Band tag reversion setup

The setup is a sequence, not a single touch:

  1. 1Plot Bollinger Bands at the defaults, a 20-period average with bands two standard deviations away, and classify the regime first: rotation back and forth across the middle band suggests a range, persistent one-sided closes suggest a trend.
  2. 2Wait for a tag: price touching or piercing an outer band, equivalent to %B reaching 1 or above, or 0 or below.
  3. 3Demand confirmation, such as a close back inside the bands, a clear rejection candle at the extreme, or an oscillator turning back from its extreme.
  4. 4Check the volatility backdrop with BandWidth: tags that occur as the bands expand out of a tight base are breakout behavior, the worst environment for fading.
  5. 5Define the trade: entry on the confirming close, stop beyond the tag extreme, first target at the middle band.

How it's calculated

A mean-reversion setup where price tags the upper or lower Bollinger Band and is traded back toward the middle band.

Middlet=1n×∑i=0n−1Ct−i\mathrm{Middle}_t = \frac{1}{n} \times \sum_{i=0}^{n-1} C_{t-i}
SDt=1n×∑i=0n−1(Ct−i−Middlet)2\mathrm{SD}_t = \sqrt{\frac{1}{n} \times \sum_{i=0}^{n-1} (C_{t-i} - \mathrm{Middle}_t)^2}
Uppert=Middlet+k×SDt\mathrm{Upper}_t = \mathrm{Middle}_t + k \times \mathrm{SD}_t
Lowert=Middlet−k×SDt\mathrm{Lower}_t = \mathrm{Middle}_t - k \times \mathrm{SD}_t
%bt=Ct−LowertUppert−Lowert\%b_t = \frac{C_t - \mathrm{Lower}_t}{\mathrm{Upper}_t - \mathrm{Lower}_t}
Lower tag (long setup): Ct≤Lowert, equivalently %bt≤0\text{Lower tag (long setup): } C_t \leq \mathrm{Lower}_t \text{, equivalently } \%b_t \leq 0
Upper tag (short setup): Ct≥Uppert, equivalently %bt≥1\text{Upper tag (short setup): } C_t \geq \mathrm{Upper}_t \text{, equivalently } \%b_t \geq 1
Reversion target: Middlet\text{Reversion target: } \mathrm{Middle}_t
C_t: close of bar t
C_(t-i): close i bars before t
Middle_t: middle band, the n-bar SMA of close
SD_t: population standard deviation of close over n bars
Upper_t: upper band at bar t
Lower_t: lower band at bar t
n: band length (commonly 20)
k: standard deviation multiplier (commonly 2)
%b_t: position of the close within the bands
i: bar offset within the window
t: bar index

Bollinger's convention is that a tag alone is not a signal; the setup pairs the tag with a confirming trigger such as a reversal bar or an indicator non-confirmation.

Population standard deviation (divide by n) is the standard band implementation; some versions define the tag with the high or low touching the band instead of the close.

In strong trends price can walk along a band, producing repeated tags without reversion.

How traders use it

  • As a range fade: when a chop or trend filter says the market is not trending, a tag of the lower band followed by a close back inside it sets up a long toward the middle band, with the mirror-image short at the upper band.
  • As a target framework: the middle band (the 20-period average by default) is the conservative reversion target, while the opposite band is the stretch target, generally reserved for well-established ranges.
  • As a filter on other signals: some systems decline fresh entries in the direction of a move that is already tagging an outer band, on the grounds that price is stretched, and wait for a pullback instead.
  • As a squeeze exclusion: tags that coincide with a TTM Squeeze release or a sharp BandWidth expansion are treated as the start of a move rather than a stretch, so reversion traders stand aside and let breakout traders have it.
  • As a risk template: because band distance already scales with volatility, many traders set stops an ATR fraction beyond the tag extreme so normal overshoot does not knock out an otherwise valid fade, and some grade the stretch with double Bollinger zones before committing.

Bollinger Band tag reversion vs. related setups

%B: %B restates price's position between the bands as a number, above 1 on an upper tag and below 0 on a lower one. The tag reversion is a setup built on that reading; %B is the measurement that makes tags screenable and comparable.

Bollinger Squeeze: The squeeze trades volatility contraction expecting expansion, a breakout premise. Tag reversion trades stretch expecting rotation, a mean-reversion premise; they are opposite uses of the same bands, separated by regime.

Keltner Channels: Keltner Channels build the envelope from ATR around an EMA, producing smoother bands that price tags more often. Fades off either can work, but standard-deviation bands react faster to volatility bursts, so a tag means a different degree of stretch.

Concept family

Volatility

57 concepts mapped · 57 in the Library

Bollinger Band Tag Reversion FAQ

Is touching the upper Bollinger Band a sell signal?

Not by itself. John Bollinger explicitly warned that band tags are not signals: in a strong uptrend price can ride the upper band for many bars. A tag only becomes a reversion candidate when the backdrop is range-bound and there is confirmation, such as a close back inside the bands or a clear rejection candle at the extreme.

What is the target for a Bollinger Band tag reversion trade?

The most common target is the middle band, the moving average at the center of the system. In well-defined ranges some traders hold for the opposite band. Neither outcome is guaranteed: tags can resolve into band walks instead of reversions, which is why stops usually sit beyond the tag extreme rather than relying on the band to hold.

What are the best Bollinger Band settings for reversion trading?

The defaults, a 20-period average with bands two standard deviations wide, are the standard starting point. Bollinger suggested nudging width with length, roughly 1.9 standard deviations for 10-period bands and 2.1 for 50-period bands, so the fraction of price contained stays similar. Tightening the bands prints more tags and more false ones.

What is a band walk and how do I avoid fading it?

A band walk is price riding an outer band through a trend, tagging it bar after bar while the move continues; Bollinger considered it a sign of strength, not stretch. The practical defenses are a regime filter that disqualifies trending conditions, a requirement that price close back inside the bands, and refusing tags that come with expanding BandWidth.

Does Bollinger Band mean reversion work in all markets?

The bands self-scale to any instrument and timeframe, but the tendency to mean-revert does not travel equally. Instruments and periods that spend more time ranging suit the fade better than persistent trenders, and the same market shifts between regimes over time. Most practitioners test the setup per market rather than assuming the edge transfers.

What confirms a tag reversion entry?

Common confirmations are a close back inside the bands, a rejection candle at the extreme, or an oscillator turning back from overbought or oversold. Bollinger's own preference was pattern-based: his W-bottoms and M-tops, adapted from Arthur Merrill's work, where the second test holds inside the band the first test pierced.

Turn Bollinger Band Tag Reversion into a trading strategy.

Take the implementation from this page into Quant, then build on it, backtest it on real data, and keep refining it with AI.