Keltner Channels
An EMA wrapped in ATR-derived envelopes. Similar idea to Bollinger Bands but built from ATR rather than standard deviation - smoother in noisy markets.
Keltner Channels on META, daily candles. Data via Financial Modeling Prep, cached server-side.
Quick reference
Keltner Channels are the volatility envelope that traders reach for when Bollinger Bands feel too noisy. Both indicators wrap a moving average in an upper and lower band. Both expand and contract with volatility. The difference is what drives the band width: Bollinger uses standard deviation, Keltner uses Average True Range.
The practical consequence: Keltner Channels react more smoothly to volatility shifts than Bollinger Bands do. A single-bar spike that throws Bollinger wide barely moves Keltner. For trend-following setups - where you want stable bands to ride along with the trend - Keltner is often the better choice.
What Keltner Channels actually measure
Keltner Channels measure price relative to a moving average envelope sized by typical bar range.
The middle line is a 20-period EMA. The upper band sits 2× ATR above it; the lower band sits 2× ATR below it. As volatility expands (ATR rises), the bands widen. As volatility contracts (ATR falls), the bands narrow. The middle EMA continues to track price.
Because ATR is built from true range (which includes gaps), Keltner adapts more gradually to volatility changes than Bollinger Bands do. Bollinger's standard deviation is computed from closes only, and a single outlier close can spike the deviation dramatically. ATR averages range across all bars in the lookback, producing a smoother volatility estimate.
The result: Keltner Channels are better for clean trend-following reads. A trending stock often "rides" the upper Keltner band the way a strong trend rides above a moving average - cleanly, without dramatic envelope swings.
The formula
Middle line = 20-period EMA(close)
Upper band = Middle line + (2 × ATR)
Lower band = Middle line - (2 × ATR)
where ATR is typically 10-period (modern setting)
The default settings have evolved over time:
- Original Chester Keltner (1960): 10-period SMA of typical price ((H+L+C)/3), bands at ± 10-period SMA of (high - low)
- Modern (Linda Bradford Raschke / Charles Le Beau): 20-EMA with ± 2 × 10-ATR bands
The modern settings are what virtually every platform defaults to today. Chester Keltner's original is rarely used.
A worked example
Suppose the 20-EMA reads at 100.00 and the 10-period ATR reads at 1.50.
Middle line = 100.00
Upper band = 100.00 + (2 × 1.50) = 103.00
Lower band = 100.00 - (2 × 1.50) = 97.00
The channel width is 6 points (103 - 97), centered on the EMA.
Now imagine volatility expands - several large-range bars in succession push ATR to 3.00. The bands re-compute:
Middle line ≈ 100.00 (EMA shifts slowly with price)
Upper band = 100.00 + (2 × 3.00) = 106.00
Lower band = 100.00 - (2 × 3.00) = 94.00
Channel width doubled to 12 points. Compare to Bollinger Bands in the same scenario: Bollinger would react more abruptly because its standard deviation is closer-coupled to single-bar moves. Keltner expanded smoothly because ATR averages range across 10 bars.
How traders actually use Keltner Channels
Three setups generate edge. The fourth (treating Keltner like Bollinger) is suboptimal because the indicators serve different roles.
1. Trend-following: riding the upper or lower band
When price closes repeatedly outside the upper Keltner band in an uptrend, the move is strong. This is the "band-ride" pattern, and Keltner shows it more cleanly than Bollinger because Keltner's smoother bands do not widen to "swallow" the trending move.
Use case: identify which instruments on a watchlist are riding the Keltner upper band as a screen for trending opportunities. Entries on pullbacks toward the middle EMA, stops below the EMA, targets at recent swing highs or upper-band touches.
2. Breakout entries
A close decisively outside the Keltner band (with rising ATR confirming volatility expansion) is a breakout signal. The expanding-channel context confirms that the move is real rather than within-noise.
Entry on the close or on a brief pullback that holds the broken band level. Stop inside the prior channel range; target a measured move of the channel's prior width.
3. The "squeeze" - Keltner inside Bollinger
The most powerful Keltner setup. Track both Keltner Channels and Bollinger Bands on the same chart. Normally, Bollinger Bands sit outside Keltner Channels (Bollinger is wider because standard deviation reacts to outliers more than ATR does).
When Bollinger contracts inside Keltner (Bollinger upper band below Keltner upper band, Bollinger lower band above Keltner lower band), volatility has compressed unusually. This is John Carter's "TTM Squeeze" - a setup that often resolves into a directional move within a few days to weeks.
The trigger: Bollinger expanding back outside Keltner. Whichever direction price is moving at that moment is the breakout direction.
The trap most retail traders fall into
The most common Keltner mistake is using it as a Bollinger substitute - applying the same mean-reversion logic (sell at upper band, buy at lower band).
Keltner is not built for mean reversion. Its smoother bands and ATR-based width are designed for trend-following. Price closing outside the Keltner band in a trend is a continuation signal, not a reversal signal. Trying to fade those closes produces the same loss pattern as fading Bollinger Bands in a trend - just slightly worse, because Keltner's smoother bands mean price stays "outside" the channel for longer.
If you want mean-reversion bands, use Bollinger and require the middle band to be flat. If you want trend-following envelopes, use Keltner and look for band-rides.
The second trap: applying the same default settings universally. The 20-EMA / 10-ATR / 2× multiplier defaults work well for daily charts on most instruments. They can be too tight on volatile instruments (crypto, small caps) and too loose on calm ones (utilities, large-cap blue chips). Adjust the multiplier (try 1.5× or 2.5×) for specific instruments if needed.
Keltner vs other volatility envelopes
vs Bollinger Bands. Bollinger uses standard deviation; Keltner uses ATR. Bollinger reacts faster to single-bar shocks; Keltner is smoother. Bollinger handles mean reversion in ranges; Keltner handles trend-following. They are complementary, not redundant.
vs Donchian Channels. Donchian is simply the highest high and lowest low over N periods - a pure breakout channel with no smoothing. Hard-edged where Keltner is smooth. Donchian is preferred for raw breakout systems; Keltner for trend-following with envelope context.
vs Chandelier Exit. Chandelier is a trailing-stop rule built from ATR (highest high since entry minus 3 × ATR). Not an envelope, but related in that both are ATR-based volatility measures. Use Chandelier for trailing stops; Keltner for envelope-based entries.
vs Moving Average Envelopes. MA envelopes are simple percentage offsets from a moving average (e.g., MA ± 2 percent). They do not adapt to volatility. Keltner is the volatility-adaptive version.
Common questions
Should I change the 20-EMA / 10-ATR / 2× defaults? Usually no. The modern defaults work well across most instruments. If you adjust, change one variable at a time - typically the ATR multiplier (1.5× for tighter signals, 2.5× for cleaner trend rides).
Why use Keltner over Bollinger? Keltner is smoother because ATR is a more stable volatility estimate than standard deviation. For trend-following, Keltner is the cleaner envelope. For mean-reversion, Bollinger is more appropriate. Many traders use both.
Do Keltner Channels work on intraday charts? Yes. The same logic applies: trend-following uses the bands as dynamic reference; the squeeze setup (Bollinger inside Keltner) works on any timeframe that has enough bars for ATR to stabilize.
What is the TTM Squeeze? John Carter's variant of the Bollinger-inside-Keltner squeeze, formalized into a single indicator with histogram. The setup is identical; the indicator just packages the comparison into one visual. Powerful when properly understood.
Can Keltner Channels predict reversals? No. They identify volatility regime, not direction. Volatility expansion is informative; direction comes from price action.
Why does Keltner sometimes look "too tight" on a chart? Because ATR is low - the market is in a calm volatility regime. The bands are doing what they should: contracting when volatility is low. The squeeze (tight Keltner) is itself a setup, not a failure of the indicator.
When to use Keltner Channels and when not to
Use Keltner Channels when:
- You are trend-following and want envelope context with smoother bands than Bollinger
- You are setting up the Bollinger-inside-Keltner squeeze trade
- You want a breakout filter that excludes single-bar volatility spikes
Skip Keltner Channels when:
- You are mean-reversion trading in a ranging market - use Bollinger Bands instead
- You are using raw breakout rules - Donchian Channels are simpler and serve that purpose better
- You already have ATR and EMA on the chart explicitly - Keltner gives you the same information packaged
