What the bands actually are, mechanically
A Bollinger Band setup has three lines. The middle line is a simple moving average, typically 20 periods. The upper and lower bands sit a fixed multiple of standard deviation away from that average, typically 2. That's the entire construction — a moving average with a statistical dispersion envelope wrapped around it.
Standard deviation here is calculated from the same lookback window as the moving average — the last 20 closes, in the default setup. When those 20 closes are tightly clustered, standard deviation is small, and the bands sit close to the middle line. When those closes are scattered — big up days, big down days, wide range — standard deviation is large, and the bands push outward.
This is the single fact that determines everything else about how the indicator behaves: the bands are a rolling measurement of how much price has been moving, not a prediction of where price is going. They react to volatility that already happened. They contain no information about direction.
The core misunderstanding
Most retail chart commentary treats Bollinger Bands as boundary lines: price hits the upper band, that's "overbought," sell. Price hits the lower band, that's "oversold," buy. This treats the bands as if they were support and resistance drawn from price levels, the way you'd draw a horizontal line at a prior swing high.
That's not what they are. Under the assumption of roughly normal distribution, price touching a 2-standard-deviation band is just... price being volatile. Nothing about that event says whether the volatility will continue in the same direction or reverse. A coin that's up 12% in three days and touches the upper band is exhibiting momentum, not necessarily exhaustion.
The bands don't know about trend, order flow, news, or exchange listings. They know one thing — the recent standard deviation of closing prices — and they draw two lines based on it.
Why band-touch selling gets run over in trends
Here's where the misuse becomes costly rather than just imprecise. In a genuine trend — a coin that's broken out on real volume and is being bid up persistently — price can ride along the upper band for extended periods. Traders call this "walking the band." Closes stay near or outside the upper band for a dozen or more candles in a row, because a strong trend keeps pushing standard deviation and price outward together.
A trader who sells every touch of the upper band during a real trend gets stopped out, or exits, repeatedly, each time watching price continue higher without them. The band gets touched, they sell, price walks another 8% up the band, touches again, they've already exited. This isn't a rare edge case — it's the normal behavior of the indicator during any sustained directional move, which is precisely when the naive interpretation fails hardest.
The same applies in reverse during a strong downtrend: price can walk the lower band for days while "oversold" buyers keep stepping in and keep getting punished. The band isn't measuring how stretched a rubber band is. It's measuring how volatile price has recently been, and volatility and trend strength are often the same phenomenon wearing different clothes.
The squeeze: the more honest signal
There is a genuinely useful pattern in Bollinger Bands, and it comes from taking the volatility-measurement nature of the indicator seriously instead of fighting it.
When the bands narrow to a multi-period low — bandwidth compressed relative to its own recent history — that's realized volatility dropping to unusually low levels. Low volatility doesn't persist indefinitely in liquid markets; it tends to resolve into a period of higher volatility. This compression-then-expansion pattern is often called "the squeeze."
The honest statistical claim is narrower than what gets repeated in chat rooms. A squeeze says: a large move is more likely soon than the recent range would suggest. It does not say which direction. Traders who treat a squeeze as a buy signal are making the same directional error as band-touch sellers, just one layer more sophisticated about it.
What a squeeze is actually useful for is positioning and risk framing — recognizing that a coin sitting in unusually tight range is a coin where a breakout, in either direction, is statistically closer than a coin still in wide, choppy price action. Some traders combine the squeeze with a separate directional filter — volume, a breakout of the recent range, or a moving-average cross — to decide which way to lean once the move starts, rather than trying to predict it in advance.
%B and bandwidth: quantifying what you're looking at
Two derived metrics turn the visual band picture into numbers, which matters if you're trying to be consistent rather than eyeballing a chart.
- %B expresses where the current price sits relative to the bands, on a scale where 0 means price is exactly at the lower band and 1 means price is exactly at the upper band. Values above 1 or below 0 mean price is trading outside the bands entirely, which happens more often than people expect during volatile stretches. %B is a cleaner way to track "how close to the edge" over time than squinting at a chart, and it makes walking-the-band episodes visible as a string of high %B readings rather than a single touch.
- Bandwidth measures the distance between the upper and lower band relative to the middle line, expressed as a percentage. This is the number that actually identifies a squeeze — plot bandwidth on its own pane and the compression shows up as a clear trough, easier to spot consistently than trying to judge "narrow" by eye on the price chart itself.
Neither metric adds new information beyond what's visually present in the bands. They just make it measurable, which is useful if you want to backtest a rule instead of reacting to a chart in the moment.
What Bollinger Bands are, stripped of the costume
Strip away the trading-room vocabulary and Bollinger Bands are a volatility indicator wearing a buy/sell-zone costume. The upper and lower lines look like boundaries, which invites people to treat them like support and resistance — trade the bounce, fade the touch. But they were never built from price structure. They were built from a moving average and a standard deviation calculation, full stop.
Most of the retail misuse traces back to that one substitution: treating a measurement of how much price has moved recently as a measurement of where price should go next. Those are different questions, and the indicator only answers the first one.
This doesn't make the indicator useless — it makes it a context tool rather than a signal generator. Knowing that a coin is currently in a low-volatility squeeze versus a wide, expanded-band trending regime changes what kind of setups make sense, what stop distance is reasonable, and how much weight to put on other signals firing at the same time. A momentum breakout signal firing while bands are still tight and bandwidth is near a multi-month low is a different situation than the same signal firing after bands are already blown wide open from a week of movement — the first has room to run, the second may already be late.
How to actually use them
Use the bands to answer "how volatile is this instrument right now, relative to its own recent history" — not "should I buy or sell at this exact price." A squeeze tells you to expect a range expansion and to size and set stops accordingly, not which direction to lean. A band touch during a strong trend tells you volatility is elevated in the direction of the trend, not that the trend is over.
Practically, that means treating bandwidth as one input feeding into position sizing and stop placement, and pairing it with a separate directional read — trend structure, volume, or another signal — before acting. If you're already using a broader scoring framework to flag setups, band width and %B are reasonable additions to the volatility-context layer of that score, not the decision by themselves. A coin with bands walking tight for two weeks and then breaking out with volume is a materially different trade than a coin bouncing off a lower band in the middle of an otherwise unremarkable range — the bands alone don't tell you which one you're looking at until you check what's driving the move.