Keltner Channels: A Volatility Band Built on ATR
A Keltner channel draws an exponential moving average, usually 20 periods, with bands set at a multiple of ATR above and below it, commonly two times.
How does it differ from Bollinger Bands? Bollinger Bands use standard deviation, which reacts quickly to sudden price spikes. Keltner Channels use ATR, which is smoother and based on the full range. The result is bands that are less jumpy.
A common use is to identify strong trends. Price closing repeatedly outside the upper band shows persistent momentum, and in strong trends it can ride the band for a long time.
Combining both indicators creates a well-known idea: when the Bollinger Bands move inside the Keltner Channels, volatility is unusually compressed. When the bands expand back outside, a move may begin. Many traders use this as a squeeze signal.
Like other bands, the channel gives context. Price between the middle line and the upper band in an uptrend suggests healthy movement. Price falling below the middle line warns of weakening.
Settings can be adjusted. A larger ATR multiple creates wider bands that trigger less often.
None of this predicts direction by itself. The bands show where price is relative to recent volatility, not where it is going.
Practical step: overlay both Keltner and Bollinger indicators on one chart and mark every time the Bollinger Bands fell inside the Keltner Channels.
Keltner Channels: A Volatility Band Built on ATR
Keltner Channels: A Volatility Band Built on ATR
It’s Fairman 
Re: Keltner Channels: A Volatility Band Built on ATR
How I'd use Keltner practically: in a trending market, a pullback to the middle EMA is often a decent spot to look for continuation. A close outside the outer band shows strength, but chasing it is usually a bad entry. One combination people like is watching Bollinger Bands move inside the Keltner Channel. When that happens, volatility is very compressed, and a breakout often follows. When the Bollinger Bands expand back outside, the move has started. It's a visual version of the squeeze idea. I'd still treat it as a heads up rather than an entry signal, and wait for structure on the chart to confirm which way price is breaking.
It’s Fairman 
-
LondonNewsTrader
- Posts: 421
- Joined: Mon Sep 21, 2026 9:30 am
Re: Keltner Channels: A Volatility Band Built on ATR
The difference shows up clearly around data releases. After a big spike, Bollinger Bands blow out on the next bar, while a 20-period ATR channel only widens gradually, so price can sit outside the Keltner band for several bars purely because the bands haven't caught up with the new volatility. Reading that as a strong trend riding the band would be a mistake.
I'd treat closes outside the channel in the first few bars after a high-impact number as noise, and only count band-riding once the ATR has had time to adjust. On a 5-minute chart that's roughly the first hour and a half.
Re: Keltner Channels: A Volatility Band Built on ATR
One difference I've noticed in practice: Keltner bands are smoother than Bollinger because ATR doesn't jump around as much. That makes them easier to read for trend pullbacks, though they're slower to show a squeeze. Some traders combine them, watching for Bollinger Bands moving inside Keltner Channels, which is the TTM squeeze idea. When that happens volatility is unusually low, and when the Bollinger Bands move back outside, the expansion has started. In a trend, a pullback to the middle EMA of the Keltner channel is a common entry zone, with the stop beyond the lower band.
It’s Fairman 
-
FXS Prop Desk
- Posts: 1689
- Joined: Wed Oct 07, 2026 9:52 am
Re: Keltner Channels: A Volatility Band Built on ATR
ATR-based bands are smoother than Bollinger Bands because ATR changes more slowly than standard deviation. That makes Keltner channels less sensitive to one large candle.
Some traders use both together: when Bollinger Bands move inside the Keltner channel, it marks a squeeze. When they expand outside it, the move has started. It is a popular combination for breakout timing.
Some traders use both together: when Bollinger Bands move inside the Keltner channel, it marks a squeeze. When they expand outside it, the move has started. It is a popular combination for breakout timing.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Keltner Channels: A Volatility Band Built on ATR
The multiplier on the Keltner side decides how often the squeeze fires, which changes the whole picture. With Bollinger at 20 periods and 2 deviations, the bands sit inside a 1.5 ATR Keltner channel only when the standard deviation drops below 0.75 of the ATR.
Widen the Keltner bands to 2 ATR and the condition becomes standard deviation below a full ATR. That happens far more often, so a squeeze that fires several times a session is often just a setting.
The lag point above works out like this. With an M5 ATR of 4 pips and a release bar with a 30 pip range, a 20 period smoothed ATR moves to 4 times 19 plus 30, divided by 20, which is 5.3.
The band at 2 ATR widens only from 8 to about 10.6 pips while price may be 30 pips away. Price can sit outside that band for a long time simply because the average has not caught up.
So it may help to note the multiplier whenever a squeeze result gets reported. A count of squeeze breakouts at 1.5 ATR and one at 2 ATR are really two different tests.
Widen the Keltner bands to 2 ATR and the condition becomes standard deviation below a full ATR. That happens far more often, so a squeeze that fires several times a session is often just a setting.
The lag point above works out like this. With an M5 ATR of 4 pips and a release bar with a 30 pip range, a 20 period smoothed ATR moves to 4 times 19 plus 30, divided by 20, which is 5.3.
The band at 2 ATR widens only from 8 to about 10.6 pips while price may be 30 pips away. Price can sit outside that band for a long time simply because the average has not caught up.
So it may help to note the multiplier whenever a squeeze result gets reported. A count of squeeze breakouts at 1.5 ATR and one at 2 ATR are really two different tests.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Keltner Channels: A Volatility Band Built on ATR
For anyone using the pullback to the middle line on a scalping timeframe, the channel width is worth comparing with the spread first. The trip from the middle EMA to the outer band is the natural target, and it shrinks with the timeframe much faster than costs do.
Say EURUSD on M1 has an ATR of 1.5 pips, so with bands at 2 ATR the band sits 3 pips from the middle line. A 0.8 pip round trip cost takes more than a quarter of that target before anything else goes wrong.
Move up to M5, where ATR might be 4 pips, and the same trip becomes 8 pips while the cost falls to a tenth. Nothing about the setup has changed except how much of it the broker keeps.
A rough filter is to require the middle to band distance to be at least eight times the round trip cost. At 0.8 pips that means 6.4 pips, which rules out most M1 sessions on the majors outside the busiest hours.
The squeeze idea runs into the same limit from the other side. A squeeze on M1 marks a quiet spell where the bands are at their narrowest, so the first part of any breakout from it is often too small to pay for the entry.
Say EURUSD on M1 has an ATR of 1.5 pips, so with bands at 2 ATR the band sits 3 pips from the middle line. A 0.8 pip round trip cost takes more than a quarter of that target before anything else goes wrong.
Move up to M5, where ATR might be 4 pips, and the same trip becomes 8 pips while the cost falls to a tenth. Nothing about the setup has changed except how much of it the broker keeps.
A rough filter is to require the middle to band distance to be at least eight times the round trip cost. At 0.8 pips that means 6.4 pips, which rules out most M1 sessions on the majors outside the busiest hours.
The squeeze idea runs into the same limit from the other side. A squeeze on M1 marks a quiet spell where the bands are at their narrowest, so the first part of any breakout from it is often too small to pay for the entry.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Keltner Channels: A Volatility Band Built on ATR
USDHKD barely moves, and that's by design. The Hong Kong dollar is pegged to the dollar inside a band of 7.75 to 7.85 per dollar. The monetary authority steps in at the edges: at 7.85 it buys Hong Kong dollars, at 7.75 it sells them. Inside the band the pair drifts with interest rate gaps and flows.
The mechanism is the interesting part. When the pair hits the weak edge, the authority buys HKD and drains bank reserves, which pushes local interest rates up and pulls money back. No dramatic announcement needed, the plumbing does the work.
Scalpers usually ignore the pair because the range is too small to cover a spread. Still, where it sits in the band can hint at funding conditions in Asia, and it's a neat example of what a defended range looks like when it isn't a hard peg.
Do you track any pegged or banded currency, and does it tell you anything about the rest of the market?
The mechanism is the interesting part. When the pair hits the weak edge, the authority buys HKD and drains bank reserves, which pushes local interest rates up and pulls money back. No dramatic announcement needed, the plumbing does the work.
Scalpers usually ignore the pair because the range is too small to cover a spread. Still, where it sits in the band can hint at funding conditions in Asia, and it's a neat example of what a defended range looks like when it isn't a hard peg.
Do you track any pegged or banded currency, and does it tell you anything about the rest of the market?
Re: Keltner Channels: A Volatility Band Built on ATR
That detail about the bands lagging after a release changes how I read it. I'd been treating several closes outside the Keltner band after a data spike as a sign of a strong trend, and then wondering why price faded. If the 20 period ATR hasn't caught up yet, the band is just too narrow for the new volatility, so "outside" doesn't mean much for the first hour or so.
On M5 I tested it on a few CPI days. After a 30 pip spike on EURUSD, the Keltner upper line at 2 ATR was about 12 pips above the mid line, while the average range per bar had briefly doubled. Price sat outside for six bars and then fell back inside. If I'd gone long on the first close beyond it with a 10 pip stop at 0.4 lots, that's 40 dollars risked on a signal that wasn't one.
My rule now is no band-based entries for 90 minutes after a big number, as you say. After that I'd trust it again, but I'd compare the ATR value with what it was just before the release.
Do you adjust the multiplier on release days, like 2.5 instead of 2, or just wait? And does the same lag affect the middle EMA line?
On M5 I tested it on a few CPI days. After a 30 pip spike on EURUSD, the Keltner upper line at 2 ATR was about 12 pips above the mid line, while the average range per bar had briefly doubled. Price sat outside for six bars and then fell back inside. If I'd gone long on the first close beyond it with a 10 pip stop at 0.4 lots, that's 40 dollars risked on a signal that wasn't one.
My rule now is no band-based entries for 90 minutes after a big number, as you say. After that I'd trust it again, but I'd compare the ATR value with what it was just before the release.
Do you adjust the multiplier on release days, like 2.5 instead of 2, or just wait? And does the same lag affect the middle EMA line?
It’s Fairman 