Bollinger Bands — Squeeze and Mean-Reversion Strategy

Bollinger Bands are one of the most popular indicators in the world — and a rare case of a tool that supports two opposite strategies at once. Some traders play mean reversion (price went beyond the band → it'll come back), others play breakout (price went beyond the band → it'll keep going). Both groups are right. And both are wrong. It all depends on which market they're doing it on — and that's not an opinion, it's a measurement result.

That makes Bollinger Bands ideal material for a "no hype" article: the same indicator, the same settings, twelve tested variants — and results that diverge depending on the asset. Anyone who doesn't know this ports a strategy from YouTube onto their own market and wonders why "it doesn't work."

What Bollinger Bands Are

The construction is simple: the middle line is a 20-period moving average (SMA20), and the upper and lower bands sit 2 standard deviations above and below it. The standard deviation is calculated from price, so the bands breathe with the market:

• volatility rises → the bands widen,
• volatility falls → the bands narrow.

At the 20/2 settings, the bands statistically contain about 95% of price movement — so a move beyond them is a rare event. What to do with that depends on the market's character:

1. Mean reversion: you treat a move beyond the band as an overreaction — you play the return to the middle. Natural in markets that "revert" (stocks, indexes).
2. Breakout (continuation): you treat a move beyond the band after a compression period as the start of a move — you play the continuation. Natural in momentum markets.
3. Squeeze: the measure of compression is BandWidth — the width of the bands (upper minus lower, ideally as a percentage of price). When BandWidth drops near the low end of its six-month range, the market is coiled like a spring: a volatility expansion is coming, though the squeeze itself doesn't say which direction.

What the Numbers Say — Measured per Market

QuantifiedStrategies tested 12 Bollinger Bands strategy variants across different markets. The result is a lesson more important than any single setting (data from QS summaries — verify at the source before citing):

Same indicator, three markets, three different answers. This actually makes structural sense: US stocks and indexes are among the most mean-reverting markets in the world (overreacted moves tend to come back more often), so fading moves beyond the band historically paid off there, while playing their continuation didn't. Gold and commodity markets trend more often — there, breakout paid off.

The practical conclusion, one that should hang over every trader's desk: you don't test a strategy "in general" — you test it on a specific market. A result measured on the S&P 500 says nothing about BTC until someone runs the numbers on BTC. Crypto has historically behaved more momentum-driven than stocks, which suggests it's closer to "gold's camp" (breakout after a squeeze) than "the stock camp" (fade) — but a suggestion isn't a measurement; test it on BTC/ETH data before you use it.

And the standing caveats: past results don't guarantee future ones, and numbers from third-party services should be verified at the source.

How to Apply It Step by Step

Variant A — squeeze + breakout (momentum markets, e.g. crypto)

1. Scan: BandWidth near the low end of its last ~6-month range; on the chart, visibly narrowed bands after a move or during a long consolidation.
2. Direction — before it breaks: the squeeze doesn't tell you the side, so you look for an edge elsewhere: level structure (which is closer — resistance or support), the higher-timeframe trend, and ideally volume indicators — accumulation on OBV or a positive Chaikin Money Flow during the consolidation raises the odds of an upside breakout; distribution points to the downside.
3. Signal: a candle closing beyond the band in the breakout direction, confirmed by a break of the consolidation level and above-average volume — a bare band break with no confirmation fails more often (the full breakout filter checklist).
4. Stop: beyond the opposite edge of the consolidation or below the breakout candle (variants); target: a projection of the consolidation's height, or a trailing stop in the trend.
5. Watch for the "head fake": John Bollinger himself warns about this trick — price breaks one way after the squeeze, reverses, and makes the real move in the opposite direction. That's why you use a candle close plus level plus volume instead of reacting to the first tick — and if a fake-out stops you out, the signal in the opposite direction remains valid.

A numerical example on ETH (illustrative): a $10,000 account, 1% risk = $100. ETH has been consolidating between $2,500 and $2,620 for a month; BandWidth is at a six-month low, OBV is rising within the range (accumulation). A daily candle closes at $2,660 — above the upper band and above the $2,620 resistance, on volume 70% above average. Entry at $2,660, stop inside the range at $2,560 (risk of $100 per ETH) → position size = 1 ETH. Target from the range projection (a $120 height): $2,740 for the first part of the position, the rest on a trailing stop. RR to the first target is about 0.8:1 — which is why the point of this play is the trend's tail, not the first target; if you don't accept managing a running position, this variant isn't for you.

Variant B — mean reversion (stock markets; on crypto, only after your own test)

1. Regime filter: the market is in an uptrend or a wide consolidation (e.g. price above the SMA200) — fading declines in a bear market is catching knives.
2. Signal: a candle closes below the lower band, followed by a candle closing back inside the bands — you play the reversion only once the market shows the overreaction is over, not on the mere touch.
3. Exit: near the average (SMA20) — the target is close, trades are short (2-6 candles), the win rate is high, individual gains are small. The mechanics closely resemble the RSI-2 system — both tools measure the same short-term overreaction phenomenon.
4. Risk: a small position size and a time stop (exit after N candles with no bounce) — in mean-reversion strategies, a classic tight stop loss usually worsened results in tests, so risk control comes from size, not a tight SL.

When It Doesn't Work and Common Pitfalls

• Fading a trend "because it touched the band." In a strong trend, price can ride the upper band for weeks — every touch is another short and another loss. A band is a statistical measure, not a wall; without a regime filter, mean reversion turns into a loss-generating machine.
• Playing the squeeze without confirming direction. A squeeze breakout can be a head fake — entering on the first tick beyond the band, without a candle close, a level and volume, systematically feeds the market your stops.
• Porting a strategy between markets without testing. This is the main lesson from the 12-variant backtest: mean reversion from stocks ported 1:1 onto a trending market (and vice versa) loses its edge. Change the market → measure again.
• Using the bands as a standalone system. In tests and in practice, the bands work best as a context frame (volatility, overreaction, compression), on top of which you layer the rest: levels, volume, trend. A bare "band touch" signal has no documented edge.
• Fiddling with parameters after a losing streak. 20/2 doesn't work, so maybe 14/1.5? Or 30/2.5? Along the way you'll find a combination that's perfect for the past and useless for the future — that's overfitting, not optimization.
• Ignoring costs in the mean-reversion variant. Short trades with a nearby target are sensitive to fees and spread — do the cost math before assuming a historical edge survives on your exchange.

One workshop tip to close with: alongside BandWidth, it's worth knowing the %B indicator, which shows price's position relative to the bands on a 0-1 scale (0 = lower band, 1 = upper band). It turns "touches" and "moves beyond the band" into a concrete, comparable number — which lets you write rules like "close below the lower band, then return inside" unambiguously and test them rigorously, instead of eyeballing them. Every strategy in this article should be expressible in numbers — if you can't write it that way, you can't test it either.

The "no hype" conclusion: Bollinger Bands are neither a dip-buying strategy nor a breakout strategy — they're a volatility thermometer that, on different markets, supports different, mutually contradictory strategies. Measurement per market (stocks: fade; gold: breakout; indexes-breakout: no edge) is the best proof in this whole series that the question "does strategy X work?" is the wrong question. The right one is: "does strategy X work on market Y — and who measured it?" If nobody has — measure it yourself, before you pay for the answer in spread and fees.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero