Bollinger Bands on ES and NQ: Band Width, the Squeeze and Walking the Band

Bollinger Bands tell you how spread out recent closes have been. They do not tell you which direction comes next. Treating the upper band as a sell signal is the most common way people get run over on a trending ES session.

The calculation

The middle band is a 20-period simple moving average of closes. The upper band is that average plus 2 times the standard deviation of the same 20 closes, and the lower band is the average minus 2 times the standard deviation. Hypothetical numbers: a 20-period average of 5,800 and a standard deviation of 12 points give an upper band of 5,824 and a lower band of 5,776. Band width, as a share of the middle band, is (5,824 minus 5,776) divided by 5,800, about 0.83 percent.

You will often read that about 95 percent of prices fall inside two standard deviations. That is true for a normal distribution, and price changes aren't normally distributed, so I wouldn't lean on the percentage. What the bands do give you is a consistent measure of recent volatility on whatever chart you use.

Walking the band

In a strong trend, price can close at or beyond the upper band repeatedly. If you sell each touch, you take a string of stops while the trend continues. A series of closes along the band shows persistent pressure in one direction, not an automatic overbought signal. If you want to fade a band touch, require a second reason, such as a failed break at a level you identified in advance, and size for the possibility that the trend simply continues.

The squeeze

A squeeze is a period when band width falls to its lowest reading in some lookback window, say the last 20 to 120 bars. Low volatility is often followed by expansion, but the direction is unknown, and some squeezes just dissolve into more chop. The squeeze identifies compression. It is not an entry.

Many charting platforms offer a TTM-style squeeze indicator, which typically compares Bollinger Bands to Keltner Channels. Availability varies, so check which version you have and read how it defines the squeeze. If you build your own filter, define it numerically: for example, band width at its lowest value in 50 bars.

Pricing a squeeze breakout

Hypothetical ES example on a 5-minute chart. The bands have been tight, the middle band is at 5,812, and price closes at 5,830 above the upper band as the bands begin to widen. If you use the middle band as the stop, the risk is 18 points. That is $900 per ES contract or $90 per MES contract. The stop is far from the entry because the middle band lags. If your daily loss limit is $500, that trade does not fit on ES at all, and you would trade MES or choose a tighter structural stop and accept more noise.

That is a cost of the setup that people skip. A wider stop means a larger target is needed for the same reward-to-risk. Measure it on your own data before you trust it.

Band width as a context reading

• Narrow and flat bands: low recent volatility, with moves that tend to be small.
• Bands starting to expand: volatility is increasing, and the direction of the first close outside the band gives a bias to test.
• Wide bands: recent moves have been large. Stops sized for a quiet tape get hit by ordinary noise, so size them to current volatility.
• Bands narrowing from wide: the large move may be fading. Decide in advance how you will manage open trades if that happens.

These readings describe conditions. They don't assign a winning approach to each state. Whatever you choose to do in each, log it and look at the results.

Band touches at planned levels

A lower-band touch that lines up with a level you planned around, plus momentum divergence, gives you a specific price to define a stop against. KPL levels are fixed reference prices anchored to the regular-session opening print at 9:30 am ET. They are not known before the open and they do not move during the session. A touch of a KPL level plus a band plus divergence is three observations, not a probability. You'll need your own logged sample to see whether stacking them helps. Thirty examples are a start, not a proven edge.

A test procedure

1. Define the squeeze numerically and the breakout as a close outside the band.
2. Mark every occurrence for 50 sessions without looking at the outcome first.
3. Record entry, stop, the point distance, commissions and slippage, and the result in R.
4. Count false breakouts separately so you can see how often price re-enters the bands.

For how KPL levels are built and used as reference points, read the KPL trading page. The YMI Intro plan is $97/month after a 7-day trial and includes the course, daily ES/NQ KPL levels and the community.