Skip to main content

Volatility Indicators (Bollinger Bands, ATR)

Updated Jun 23, 2026 at 6:04pm

Research Draft High 722 words

Volatility indicators measure how much price is moving, not which way. They don't generate direction — they size the move, frame what "normal" looks like right now, and (critically for risk) tell you how far a stop has to sit to survive noise. The two workhorses are Bollinger Bands (relative volatility around a mean) and ATR (absolute volatility of the bar). Both are inputs to sizing and stops, not buy/sell signals on their own.

Bollinger Bands

Originator: John Bollinger (early 1980s). Three lines around price:

  • Middle band: 20-period SMA (default)
  • Upper band: middle + (2 × 20-period standard deviation)
  • Lower band: middle − (2 × 20-period standard deviation)

Defaults are 20, 2. The period and the deviation multiple are coupled — Bollinger recommends ~2.1σ for a 50-period band and ~1.9σ for a 10-period band, so the bands stay statistically meaningful when you change the lookback.

Two derived tools:

  • %B = (close − lower) / (upper − lower) — where price sits in the channel: 1.0 = upper band, 0.5 = middle, 0.0 = lower, and outside [0,1] when price is beyond a band.
  • BandWidth = (upper − lower) / middle × 100 — normalized width; the basis of the squeeze.

How it's read:

  • The Squeeze: historically low BandWidth = volatility contraction, which tends to precede a large directional move (direction unspecified — the squeeze doesn't say up or down).
  • Walking the bands: in a strong trend price rides one band for an extended run — a sign of sustained momentum, not a reversal cue.
  • M-tops / W-bottoms: reversal structures Bollinger confirms with band position (e.g. a lower high that still tags the band vs one that doesn't).

The critical caveat (Bollinger's own rule): "Tags of the bands are tags, not signals." Touching the upper band is not a sell, touching the lower band is not a buy — in a trend, price can ride a band for weeks and punish every fade. A close outside a band is initially a continuation signal, not a reversal. In practice ~85–90% of price action stays inside the default bands (less than the 95% a normal distribution implies — markets have fat tails). Bands confirm; they don't trigger.

ATR (Average True Range)

Originator: J. Welles Wilder (1978). ATR measures absolute volatility per bar.

True Range is the max of three values, so it captures overnight gaps that a simple high−low misses:

TR = max[ (High − Low), |High − prev Close|, |Low − prev Close| ]

Example: High 152.40, Low 148.60, prior Close 147.80 → candidates 3.80, 4.60, 0.80 → TR = 4.60 (the gap from the prior close, not the intraday range, is the true move).

Averaging (Wilder's smoothing): default period 14. The first ATR is a simple average of the first 14 TRs; thereafter ATR_today = ((ATR_prev × 13) + TR_today) / 14 — an exponential smoothing with factor α = 1/n (slower than a standard EMA's 2/(n+1)), so a single spike bleeds in gradually.

How it's used (all non-directional):

  • Stops: volatility-scaled stops — e.g. a chandelier exit at close − 3 × ATR (long) — so the stop sits beyond normal noise and widens/tightens with conditions.
  • Position sizing: size inversely to ATR — higher ATR → smaller position for the same dollar risk. This is the backbone of risk-per-trade math.
  • Breakout context: contracting ATR marks coiled consolidation; an ATR expansion can confirm a breakout has real participation.

Caveats / failure modes:

  • Absolute, not percentage. ATR $0.50 on a $10 stock (5%) is far more volatile than ATR $2.00 on a $200 stock (1%). For cross-name comparison use ATR% = ATR / close × 100.
  • Direction-blind. ATR rises in both up- and down-moves — a rising ATR is not bullish; it just means bigger bars.
  • Lagging. It's an average of past bars; it spikes after a shock, it doesn't predict one.
  • Period- and gap-sensitive. Different lookbacks give materially different values; large gaps still distort it temporarily.

Sources

  • John Bollinger — Bollinger on Bollinger Bands and the official "Bollinger Band Rules" (bollingerbands.com)
  • StockCharts ChartSchool — Bollinger Bands, Bollinger BandWidth, %B, Average True Range
  • J. Welles Wilder, New Concepts in Technical Trading Systems (1978) — True Range, ATR, Wilder smoothing
  • Investopedia / Wikipedia / Fidelity / Macroption — cross-checks on formulas, defaults, smoothing
  • Cross-source note: "% of price within default bands" verified at ~85–90% (not the theoretical 95%), reflecting fat tails