OptiNod Academy

Bollinger Bands: Volatility Channels

Use the frequency of closes outside the bands and the band-width percentile to judge current volatility. A band touch alone is not enough.

If you use Bollinger Bands as a sell signal at the upper band and a buy signal at the lower band, a trending market will keep putting you on the wrong side of the trade. The first questions are whether price can close outside the bands, and where the current band width sits within its recent distribution.

The absolute value of band width, such as 0.04 or 0.08, means different things for different assets. So compare it with the percentile rank over the prior six months. That tells you whether volatility is high or low right now. The 5th percentile usually indicates compression; the 95th percentile usually indicates an overheated zone.

Whether the high merely touched the band is not very important. It may have been only a brief probe. The real decision comes from where the candle closes. If only the upper wick touches the upper band and the candle closes back inside the band, it does not mean much. But if the candle closes outside the band and the next candle holds that area, treat it as a point where volatility may be changing.

**A touch and a close outside the band are different**A brief high above the band does not mean much. What matters for judging a volatility shift is whether the close remains outside the band.

Frequency of Closes Outside the Bands: One Number That Defines the Current Regime

Count how many of the last 20 candles closed outside the bands. Under the statistical definition of a 2σ band, only about 5% of closes should be outside the bands, so on a 20-candle lookback, one candle is the baseline. This single-digit number helps classify the current market regime.

0-1 candle: Normal volatility. Pattern setups such as W-bottoms and M-tops tend to work cleanly.
2-3 candles: Volatility is expanding. Stop using simple mean reversion and switch to trend-following tools. Watch whether the same asset returns to normal volatility within the next few days.
4 or more candles: Volatility has stepped up, or price is walking the band. If closes keep printing outside the band in one direction, it may be the start of a new trend or the middle of a strong one. Countertrend entries repeatedly end in stop-outs.

In April 2024, DXY closed outside the upper band on seven daily candles and walked the band higher from 105.5 to 106.5. Traders who kept trying to buy lower-band touches had to re-enter at higher prices each time. If they had checked close frequency first, this was the point where the method should have shifted to trend following.

That is why this one number acts as a switch between mean reversion and trend following. If the frequency is one candle or less, mean reversion stays on. If it is two candles or more, turn it off.

**Use outside-band close frequency to choose the trading mode**Counting 0-1, 2-3, or 4-plus outside-band closes over the last 20 candles helps decide whether to use mean reversion or trend following.

W-Bottom: The Second Low Is Shallower Than the First

The shape of a W-bottom is simple. Near the end of a downtrend, price makes a new low, which is the first low. It then rebounds and later makes another low, the second low. The key is that the second low breaks the lower band less deeply. On a candle chart, the structure looks like a W, but the important part is its relationship with the bands. The first low is below the lower band; the second low is inside or slightly above the lower band.

The meaning of this pattern goes deeper than its shape. The first low forms when volatility has expanded sharply. The second low revisits the same price area, but with lower volatility. Selling pressure has likely passed its peak, and the force needed to drive price back down to that area has weakened.

Entry comes after the second low forms. The trigger is the candle that closes above the interim rebound high between the first and second lows. That candle forms the upper-right corner of the W.

> After EURUSD declines for four weeks on the daily chart,

> the first low closes below the lower band, then price rebounds to the middle band over the next 10 candles.

> On the next pullback, the second low closes inside or slightly above the lower band.

> Enter long at the close of the candle that breaks above the interim rebound high on a closing basis.

> Place the stop below the second low.

> If price closes below the second low, the setup is invalid.

The key point is that the second low does not break the lower band as deeply as the first. Even if both lows form around the same price area, the second low stops inside the band because volatility has contracted. This is why Bollinger emphasized this pattern so strongly.

An M-top is the exact opposite. The second high fails to break the upper band as clearly as the first. Entry comes when price closes below the interim pullback low between the two highs.

**A W-bottom measures the strength of the second low**When the second low holds inside the band and momentum improves, it confirms that selling pressure is fading.

The Momentum Confirmation Bollinger Told Traders to Use With W and M Patterns

Bollinger made this point clearly in his book, but it is often overlooked. At the second low of a W-bottom, momentum indicators such as RSI and MFI should be higher than they were at the first low. Price has returned to the same area, but momentum has fallen less. That is the second piece of evidence that selling pressure has been exhausted. In an M-top, the opposite applies: momentum at the second high should be lower than at the first high.

The reason Bollinger recommended using MFI, or Money Flow Index, alongside RSI is straightforward. MFI is RSI with volume included, so it can confirm when price and volume fail to make a new peak together. When both RSI divergence and MFI divergence appear with a W or M pattern, the setup becomes more reliable.

When EURUSD formed a W near 1.0600 in September 2024, RSI at the second low was 8 points higher than at the first low, and MFI also failed to make a lower low. Price then continued higher to 1.1200. By contrast, a W without momentum confirmation often ends as nothing more than a basic double bottom.

After a Squeeze, Volume on the First Big Candle Separates Real Breakouts From False Ones

A squeeze with band width at or below the 5th percentile means a large move is likely coming soon. It does not tell you the direction in advance. The entry signal comes after the breakout direction is set. The trigger is a candle that closes outside the band.

This trigger has one additional condition. The candle's volume must be at least 1.5 times the average of the prior 20 candles. For assets that trade 24 hours a day, such as FX, compare volume against the same time window instead, using the average for the prior 20 days in the same GMT time slot. Without volume confirmation, a quiet close outside the upper band has a 50% or higher chance of returning inside the band within the next one or two candles.

In July 2024, DXY's band width fell to the 3rd percentile of its 100-candle distribution. On the first candle that closed outside the upper band, volume for that time window was 2.1 times normal. DXY then continued higher for six weeks. By contrast, in EURUSD in November of the same year, price closed outside the upper band after a squeeze without volume confirmation, then returned inside the band within the next two candles. It was a false breakout.

**Confirm squeeze breakouts with volume**After band-width compression, the first close outside the band should come with volume above the prior average to reduce false breakouts.

2σ Gets Broken Often

Bollinger Bands assume a normal distribution, but financial time series do not follow a normal distribution. They have *fat tails*, meaning 2.5σ and 3σ events can occur several times in a single year. If you treat 2σ as a line price cannot cross, losses can build quickly during fat-tail events.

There are two ways to respond.

Cut simple mean-reversion size to half of normal: Position size should reflect the fact that one fat-tail event can wipe out the profit from five ordinary trades.
Use the second close outside 2σ as the cutoff for mean-reversion entries: One 2σ break can be noise, but two consecutive closes outside 2σ signal that volatility has changed. When the second close appears, stop taking mean-reversion entries immediately.

For high-volatility assets such as commodities, it is better to plot a 2.5σ band as an auxiliary line. In assets such as natural gas, 2σ mean-reversion signals appear too often, so the 2σ band does not separate regimes well. Adding a 2.5σ line helps distinguish true extremes from normal volatility.

**Closes outside 2σ happen often**Because financial time series have fat tails, automatically fading every 2σ break is dangerous.

Three Mistakes That Break Simple Mean Reversion

Changing the MA type casually: Bollinger himself recommends SMA(20). If you switch to EMA, the same data produces different signals, and backtest results can diverge from live trading. Keep it fixed at SMA(20).
Using W and M patterns on short timeframes: W and M patterns appear often on 5-minute and 15-minute charts, but their reliability is low. In FX and commodities, W and M patterns become more trustworthy from the 4-hour chart upward. In stocks and crypto, start with the daily chart.
Using absolute numerical thresholds: A rule such as "band width of 0.04 is narrow" breaks down from one asset to another. Convert every judgment to a percentile. Once the percentile framework is established, the same standard can carry across assets.

Where Signals Overlap With Other Momentum Tools

The combination Bollinger emphasized in his book was momentum confirmation for W and M patterns. But once you also use the close-frequency rule, another overlap becomes useful. When frequency rises to 2 or more, check whether RSI or MFI is also crossing the 50 line in the same direction.

If both signals point in the same direction on the same candle, volatility is changing and momentum is aligned with it. That lowers the chance that the move is just a one-candle pop. Frequency alone can pull you into false moves a little early, but waiting for momentum confirmation can reduce the false-signal rate even if it means entering one candle later.

**Use frequency and momentum together**Outside-band close frequency alone can be early. Trend-shift confidence improves when momentum tools such as RSI and MFI point in the same direction.