Keltner Channel

An EMA wrapped in ATR-scaled bands. Mechanically similar to Bollinger Bands but with two crucial differences: the middle line is an EMA (not SMA), and the band width comes from ATR (not standard deviation). The result is smoother, more trend-friendly bands.

What it measures

• Volatility envelope (ATR-based). Bands expand and contract with ATR — true range, not standard deviation.
• Trend reference (EMA-based). The middle line is an EMA, which reacts faster than Bollinger's SMA and produces smoother lines in trends.
• Channel breakouts. Price closing outside the bands is the primary trade signal — more meaningful in Keltner than in Bollinger because Keltner's bands don't expand as dramatically during noise.

Formula

Middle = EMA(price, basis_period)   (default basis: EMA, period: 20)
Volatility = ATR(volatility_period) (default period: 10, Wilder smoothing)
Upper = Middle + multiplier × Volatility
Lower = Middle − multiplier × Volatility

Inputs in AlgoLift

Recommended settings

• (EMA 20, ATR 10, 2.0×) — default: Raschke's refinement. Standard for most strategies.
• (EMA 20, ATR 14, 1.5×): Tighter bands — for shorter holding periods.
• (EMA 50, ATR 20, 2.5×): Wider, slower — for swing setups.

Outputs in AlgoLift

How to read it

• Price riding the upper band: Strong uptrend in progress. In Keltner this is normal (less squeeze pressure than Bollinger), not a reversal warning.
• Price riding the lower band: Strong downtrend.
• Price oscillating between bands: Range-bound conditions.
• Price closing above the upper band: Breakout signal — Keltner's more conservative band-touching makes this more meaningful than the same Bollinger event.
• Middle line slope: The EMA-based middle gives a smooth trend reading. Rising = bullish bias, falling = bearish.

Best in / worst in

Best in trending markets and on instruments where you want clean breakout signals. Particularly good when paired with directional confirmation indicators.

Worst in very-choppy sideways markets (price oscillates within the bands continuously) and on very-low-timeframe charts where the EMA-ATR combination creates lag.

Three setups

1. Channel breakout entry

The signature Keltner pattern — trade price closing outside the bands.

• Long: Price closes above Upper Band AND Middle Line slope > 0 (trend confirmation).
• Short: Price closes below Lower Band AND Middle Line slope < 0.

The slope check filters out countertrend breakouts that often fail.

2. Pullback to the middle in trend

In trending markets, use the middle line as a dynamic pullback target.

• Long pullback: Middle Line slope > 0 (uptrend) AND price closes within 0.3 × ATR of the Middle Line AND price > Middle Line after the touch.
• Short pullback: Symmetric.

Better R:R than chasing breakouts because you're entering close to the trend's mean.

3. Bollinger / Keltner confluence (the "TTM Squeeze")

A famous technique pairing the two band systems.

• When Bollinger Upper is inside Keltner Upper AND Bollinger Lower is inside Keltner Lower, volatility has compressed unusually — Bollinger's σ-based bands have contracted faster than Keltner's ATR-based bands.
• This is the "squeeze" signal — pent-up volatility, often precedes a directional breakout.
• Enter on the first close outside Keltner's bands after the squeeze.

The two indicators measure volatility differently, so when they diverge, the difference itself is information.

Advanced patterns in AlgoLift

Asymmetric bands for directional bias. Enable Separate Multipliers to set upper multiplier = 2.5 and lower multiplier = 1.5 (or vice versa). Use for instruments with persistent skew — equity indices in bull regimes, where upside moves are larger than downside corrections.

Keltner Middle Line as a trailing stop. In a long position, exit when price closes below the Middle Line. The ATR-based smoothing gives a stop that adapts to volatility without overreacting to single bars. Cleaner than fixed-distance stops, less aggressive than Parabolic SAR.

Width compression for breakout timing. Apply ROC Modifier to Channel Width — when ROC is strongly negative (width contracting), the channel is compressing. Set a Set User Variable flag and arm a breakout entry that fires on the first close outside the bands.

Multi-timeframe Keltner regime detection. Daily Keltner gives broad regime context (price-relative-to-bands on the daily); 5m Keltner gives entry timing. Only allow 5m breakouts in the direction of the daily band position. See the multi-series Visual Builder pattern.

Common mistakes

• Treating Keltner as a clone of Bollinger. They produce different signals — sometimes meaningfully different. The two indicators are complementary, not redundant.
• Using Keltner for mean reversion. Keltner's design favors trends. Bollinger is the better choice for mean-reversion fading; Keltner is better for breakout following.
• Tuning the multiplier too low. Keltner with multiplier 1.0 becomes a noise filter — too many touches. Stay at 1.5–2.5 unless testing a specific use case.
• Ignoring the basis type option. Most users default to EMA without considering SMA. SMA basis produces a Bollinger-like middle but with ATR-based bands — sometimes useful when you want stable middle but volatility-aware width.