Bollinger %B

A single number that states where price sits relative to the Bollinger Bands: 1 at the upper band, 0 at the lower, and beyond either when price escapes the bands.

What it is

Bollinger %B, developed by John Bollinger as a companion to his Bollinger Bands, converts price's position within the bands into one normalized value. A reading of 1 means price is exactly at the upper band, 0 means exactly at the lower band, and 0.5 means at the middle band (the moving average). Crucially — and this answers the most common question about the tool — %B is not bounded: the bands do not contain price, so %B above 1 simply means price has closed above the upper band (more than the chosen number of standard deviations above the rolling mean), and %B below 0 means price has closed below the lower band. Neither reading is an error, and neither is a reversal signal; in strong trends price can ride outside a band while %B holds above 1 or below 0 for multiple bars. %B is a descriptive normalization of information already visible in the bands. It predicts nothing, we make no claim that any %B level implies mean reversion, and this entry is educational material, not financial advice — trading involves risk of loss.

How it works

Standard Bollinger Bands are built from a 20-period simple moving average of price (the basis) plus and minus 2 rolling standard deviations of the same price series. %B then measures where price falls within that envelope: %B = (price − lower band) / (upper band − lower band). When price equals the basis, the numerator is exactly half the denominator and %B reads 0.5. Because the standard deviation is recalculated each bar from the trailing window, both the band width and %B respond to recent volatility: the same absolute price move produces a larger %B change when the bands are tight than when they are wide. That adaptivity is the tool's point, but it has a mechanical consequence worth stating plainly — in a low-volatility squeeze the denominator becomes small, so %B gets hypersensitive and can swing from near 0 to near 1 on trivial price movement. A related honesty note: the '2 standard deviations' construction borrows the language of the normal distribution, but financial returns are not normally distributed, so %B above 1 must not be read as a statistically calibrated statement like 'a 5% tail event'. It means only that price is unusually far above its own recent rolling mean, by that window's own recent yardstick.

How traders read it

• Fixed references: 1 = at the upper band, 0.5 = at the basis (the 20-period average), 0 = at the lower band; between those, %B is a linear position within the envelope.
• Above 1 / below 0: price has closed outside the bands — a statement that the move is large relative to recent volatility, not an overbought/oversold trigger; in strong trends %B can stay outside 0–1 repeatedly.
• Walking the band: persistent readings in the 0.8–1.0+ zone (or 0.0–0.2 and below) describe price hugging one band, which is characteristic of trending conditions rather than an imminent turn.
• Midline behaviour: some traders watch whether pullbacks hold above 0.5 in uptrends (or below 0.5 in downtrends) as a normalized version of watching price against its moving average.
• Divergence: price making a new high while %B makes a lower high says the second high sits less far outside recent volatility than the first — context about relative stretch, unreliable as a timing tool.
• Pair with BandWidth: %B tells you where price is inside the envelope; Bollinger BandWidth tells you how wide the envelope is — %B readings taken during a tight squeeze deserve extra skepticism.

Common settings

The defaults inherit from standard Bollinger Bands: length 20, multiplier 2.0, source close. Longer lengths or larger multipliers make excursions beyond 0 and 1 rarer; shorter or tighter settings make them routine. Any change to the underlying band settings changes every %B value, so the settings should always be stated alongside readings.

Strengths

• Reduces a three-line band overlay to one comparable number, making price-vs-band position easy to reference precisely ('%B = 0.94') instead of by eye.
• Normalized output is comparable across instruments and timeframes in a way raw price-to-band distance is not.
• Adapts automatically to volatility because the bands themselves do — the same tool works across quiet and volatile regimes without re-scaling.
• Simple, fully deterministic formula that is easy to implement and verify against the bands on the chart.
• Makes band-walking visible and countable, which is a useful honest corrective to the 'touch the band, expect reversion' folklore.

Pitfalls to watch

• The mean-reversion fallacy: %B above 1 is not a sell signal and below 0 is not a buy signal — price can walk along or outside a band for extended stretches in trends, and we make no claim that any %B level implies a snap-back.
• Squeeze hypersensitivity: when the bands are tight the denominator is small, so tiny price moves produce large %B swings; readings during low-volatility periods are mechanically noisy.
• Borrowed statistics: the 2-standard-deviation construction does not mean %B outside 0–1 is a calibrated probability statement — returns are fat-tailed and non-normal, so 'rare' by Gaussian logic is not rare in markets.
• Settings sensitivity: every value depends on the band length and multiplier; the same bar can read 1.05 on one setting and 0.85 on another.
• Inherited lag: the basis and standard deviation are rolling calculations over past bars, so the reference frame %B measures against always trails current conditions.
• No predictive power: %B restates where price sits relative to its own recent volatility envelope; it forecasts nothing, and no indicator removes the risk of loss.

Pine v6 example

//@version=6
indicator("Bollinger %B Example", overlay = false)

length = input.int(20, "Length", minval = 1)
mult   = input.float(2.0, "StdDev Multiplier", minval = 0.1, step = 0.1)
src    = input.source(close, "Source")

basis = ta.sma(src, length)
dev   = mult * ta.stdev(src, length)
upper = basis + dev
lower = basis - dev

// %B = position of price within the bands (can exceed 1 or drop below 0)
pctB = (src - lower) / (upper - lower)

plot(pctB, "%B", color = color.blue)
hline(1.0, "Upper Band (1.0)", color = color.gray)
hline(0.5, "Basis (0.5)", color = color.new(color.gray, 50))
hline(0.0, "Lower Band (0.0)", color = color.gray)

Pro tip: Always read %B together with band width: a %B of 1.1 during expanding, wide bands describes a genuinely strong move, while the same 1.1 during a tight squeeze may reflect a few pips against a tiny denominator. Logging how often your instrument's %B exceeds 1 or drops below 0 is also a quick, honest way to see that band breaches are routine events, not rarities. Educational context only — %B is descriptive, implies no mean reversion, and all trading carries risk of loss.

Educational only — not financial advice, not a recommendation to trade. No indicator is predictive; trading involves substantial risk of loss.