NVDA closed at $149.38 on 6 January 2025 with a daily range of $4.34 (high $152.11, low $147.77). The next day, the range exploded to $13.12 (high $153.08, low $139.96), and the stock closed at $140.09. The Keltner Channels would have been tight going into that session and then snapped wide open as ATR surged. If you were watching Keltner Channel Width, you would have seen the expansion starting in real time as the channel ballooned on that single session.
Keltner Channel Width (KCW) measures the distance between the upper and lower Keltner Channels as a percentage of the middle line. It is a pure volatility gauge. When the channels contract, KCW drops. When they expand, KCW rises. Low KCW means the market is quiet and a move is building. High KCW means volatility has already arrived.
I prefer KCW over Bollinger Band Width for one reason: it uses the Average True Range instead of standard deviation. ATR smooths volatility measurement across gaps and wicks. Standard deviation reacts sharply to a single outlier candle. For swing trading, where I need a stable read on whether volatility is compressing or expanding, KCW gives me cleaner signals.
The Formula
The Keltner Channel has three components: a middle line (typically a 20-period EMA), an upper band, and a lower band. The bands are set at a fixed ATR multiple above and below the middle line:
\text{Upper KC} = \text{EMA}(20) + m \times \text{ATR}(10) \text{Lower KC} = \text{EMA}(20) - m \times \text{ATR}(10)where m is the multiplier (typically 2).
Keltner Channel Width is the distance between the bands expressed as a percentage of the middle line:
\text{KCW} = \frac{\text{Upper KC} - \text{Lower KC}}{\text{Middle KC}} \times 100Since the upper band is \text{EMA} + m \times \text{ATR} and the lower band is \text{EMA} - m \times \text{ATR}, the numerator simplifies:
\text{KCW} = \frac{2m \times \text{ATR}(10)}{\text{EMA}(20)} \times 100With the standard multiplier of 2, this becomes:
\text{KCW} = \frac{4 \times \text{ATR}(10)}{\text{EMA}(20)} \times 100This tells you something useful: KCW is essentially a normalized ATR. It measures the Average True Range as a percentage of the moving average price. A KCW of 8 means the Keltner Channel spans 8% of the current price. A KCW of 3 means it spans 3%. The absolute values vary by instrument, but the relative changes (contracting or expanding) are what drive trading decisions.
What KCW Tells You
KCW is a one-dimensional indicator. It measures one thing: how wide the Keltner Channel is relative to price. It does not tell you direction. It does not tell you whether the next move will be up or down. It tells you whether volatility is low, high, or changing.
When KCW is at its lowest levels in the past 50-100 sessions, volatility is compressed. The market is coiled. A directional move is likely approaching, though KCW cannot tell you which direction.
When KCW is rising sharply, volatility is expanding. A move is already underway. Rising KCW confirms that the breakout is real and backed by genuine range expansion, not just a wick or a gap that will fill.
When KCW is falling from elevated levels, volatility is contracting after a move. The trend may be losing steam, or the market may be consolidating before the next leg. Falling KCW does not mean the trend is over. It means the urgency is fading.
The Squeeze: KCW at Multi-Week Lows
The primary trading application of KCW is squeeze detection. A squeeze occurs when KCW drops to its lowest levels in 40-60 sessions, indicating that the Keltner Channels have compressed to an unusually narrow width. This means the stock’s daily ranges have shrunk relative to its price, and a range expansion is statistically likely.
AAPL from late January into early February 2025 showed a clean compression. After the volatility of 31 January (a $13.68 range day, with a high of $245.85 and a low of $232.17), daily ranges contracted dramatically. By 6 February, the daily range had narrowed to $3.35 (high $232.53, low $229.18). The Keltner Channels would have been widening on 31 January as ATR spiked, then progressively tightening through 3-6 February as ATR decayed back toward its mean. KCW would have shown a clear decline from the January spike, approaching the pre-event compression levels. That tightening is the setup. The next expansion out of that compression is the trade.
The squeeze itself is not the signal. It is the context. You need a directional trigger (a close outside the upper or lower band, a momentum shift, or a price pattern) to decide which way to trade. KCW tells you the spring is coiled. Something else tells you which direction it uncoils.
Expansion Confirmation
Once a breakout occurs, rising KCW confirms that volatility is expanding to support the move. This is the difference between a real breakout and a fakeout: real breakouts are accompanied by range expansion. The channels widen because the stock is covering more ground per session.
NVDA from 6 to 13 January 2025 showed a sustained expansion. On 6 January, the daily range was $4.34 (high $152.11, low $147.77). On 7 January, it tripled to $13.12 (high $153.08, low $139.96). Over the next three sessions, ranges remained elevated: $6.39 on 8 January (high $143.90, low $137.51), $5.69 on 10 January (high $139.87, low $134.18), and $3.98 on 13 January (high $133.45, low $129.47). KCW would have surged on 7 January and remained elevated through 10-13 January. The widening channels confirmed that the breakdown from $149 was a genuine volatility event, not a one-day wick.
When price breaks out of a squeeze but KCW does not expand (the channels stay narrow), be cautious. The range has not confirmed the move. This often precedes a quick reversal back inside the range.
How KCW Differs from Bollinger Band Width
Bollinger Band Width (BBW) and Keltner Channel Width measure the same concept (volatility compression and expansion) but using different engines. The difference matters in practice.
BBW uses standard deviation, which is sensitive to outliers. A single candle with an unusually large range can spike standard deviation and blow the Bollinger Bands wide open. Two sessions later, when the outlier drops out of the lookback window, the bands snap back. This creates erratic readings on BBW.
KCW uses ATR, which is calculated using an exponential moving average (or simple moving average, depending on the implementation) of true ranges. ATR rises and falls more gradually because it is an average, not a single-period calculation. A large range day increases ATR, but not as dramatically as it affects standard deviation. And it decays smoothly rather than falling off a cliff when the outlier exits the window.
In practice, this means KCW gives you a smoother volatility curve. Squeezes on KCW tend to develop over multiple sessions and resolve gradually. Squeezes on BBW can appear and disappear quickly because a single volatile session distorts the reading. I use both, but when they disagree, I give more weight to KCW for identifying genuine compressions.
The classic squeeze trade (Bollinger Bands inside Keltner Channels) uses both indicators together. When the Bollinger Bands narrow enough to fit inside the Keltner Channels, you have a dual-confirmed compression. KCW is the tool that tells you how wide the Keltner side of that comparison is.
Trading the KCW Cycle
Volatility is cyclical. It contracts, then expands, then contracts again. KCW makes this cycle visible.
The trading cycle works like this: KCW trends lower over multiple sessions as daily ranges shrink. At some point, KCW reaches a level that is low relative to its own history (the lowest reading in 40-60 sessions is a useful benchmark). The squeeze is on. Then a catalyst arrives (earnings, a news event, a sector rotation, or simply accumulated supply-demand imbalance), and KCW begins rising as ranges expand. The expansion phase lasts until the move exhausts itself, and KCW peaks and begins declining as ranges normalize.
The highest-probability entry is at the start of the expansion phase: KCW has been low and is now turning higher, and price has closed outside the Keltner Channel in the direction of the breakout. The highest-probability exit (or profit target) is when KCW peaks and begins declining after a sustained expansion, suggesting the move is losing momentum.
SPY from 6 to 21 January 2025 illustrated the full cycle. From 6-8 January, ranges were elevated (SPY covered $6.02, $10.81, and $5.31 on those sessions). By 13-16 January, ranges had stabilized ($6.31, $6.56, $4.67, $3.37). KCW would have been declining through mid-January as ATR pulled back. Then on 17 January, the range widened slightly ($3.70) and by 21 January the market was trending higher with a $4.33 range. The contraction phase from 8-16 January set up the next directional move. Watching KCW bottom and turn higher was the confirmation to re-engage.
Settings and Defaults
The standard Keltner Channel settings are a 20-period EMA for the middle line, a 10-period ATR, and a multiplier of 2. These produce KCW readings that work well on daily charts for swing trading.
Shorter ATR periods (5-7) make KCW more responsive. Longer ATR periods (14-20) make it smoother. I prefer the 10-period ATR because it balances responsiveness and smoothness. The 20-period EMA is standard and I have not found a compelling reason to change it.
The multiplier affects the absolute KCW value but not the relative changes. A multiplier of 1.5 produces narrower channels and lower KCW readings. A multiplier of 3 produces wider channels and higher readings. The squeeze and expansion patterns look the same regardless of multiplier. I stick with 2 because it is the default on most platforms and ensures consistency when comparing across instruments.
For the squeeze lookback (how far back to check whether KCW is at a low), 50-60 sessions works for daily charts. This covers roughly one quarter of trading. Shorter lookbacks (20-30) produce too many false squeeze signals. Longer lookbacks (100+) miss intermediate compressions that can produce tradeable moves.
Where KCW Falls Short
KCW does not tell you direction. A low KCW reading does not mean “buy.” It means “volatility is compressed.” The next move could be up or down. You need a directional tool (trend, momentum, price action) to determine which side to trade.
KCW can stay compressed for a long time. Some stocks grind sideways for weeks with slowly declining KCW. The squeeze is real, but the catalyst does not arrive. Trading the squeeze too early (before expansion begins) is a common mistake. Wait for KCW to turn higher, not just to reach a low.
In strongly trending markets, KCW may remain elevated for extended periods. It does not mean the trend is about to reverse. It means the stock is covering a lot of ground. Trying to use high KCW as a sell signal in a strong trend will take you out of winners too early.
KCW is also less useful on instruments with erratic volume patterns. Penny stocks, thinly traded ETFs, and some commodity futures can produce noisy ATR readings that distort KCW. Stick to liquid instruments where the ATR reflects genuine price movement, not random gaps on low volume.
Putting KCW to Work on Your Charts
Not every platform has a dedicated Keltner Channel Width indicator. If yours does not, you can build it. Plot the Keltner Channel with standard settings (20 EMA, 10 ATR, multiplier 2). Then create a custom study that calculates (Upper Band – Lower Band) / Middle Band * 100. Most platforms with a custom formula editor can handle this in one line.
Once you have KCW plotted in a subpanel below your price chart, the visual pattern is immediate. The line slopes down during compressions and slopes up during expansions. Mark the lowest point on a rolling 50-session basis. When KCW dips to that level, you have a squeeze candidate. When it turns higher, you have expansion confirmation.
I run KCW alongside Bollinger Band Width on the same chart. When both are at multi-week lows simultaneously, the squeeze signal is stronger. When they diverge (BBW is expanding but KCW is flat, or vice versa), I trust KCW more for the reasons explained above. The smoother read wins for identifying genuine compressions versus noise.
Educational content only. Not investment advice. Trading involves risk. You are responsible for your decisions.