On 6 October 2026, SPY (the S&P 500 exchange-traded fund, or ETF) closed at $779.09. That was just above its upper Bollinger Band, at $778.83. Fourteen trading days earlier, on 16 September, it had closed below its lower band.
Bollinger Bands draw a channel around a price. The channel widens when the price swings more. A textbook rule reads a close below the lower band as a buy signal, while a close above the upper band is a sell signal. This guide works out the bands by hand on SPY, then tests that rule on 15 funds, the oldest from 1993.
What are Bollinger Bands? Bollinger Bands are three lines drawn around a price: a 20-day average in the middle, with an upper and a lower band two standard deviations away. A standard deviation measures how far prices typically sit from their average, so the bands widen when prices swing more.
Since 1993, SPY's median gain in the month after a close below its lower band was 2.21%, against 1.43% after all days. A close counts as a signal only when the day before was still inside the band.
After a close above its upper band, SPY rose in 64.0% of months, against 65.7% after all days.
Across 15 funds, 11.3% of daily closes fell outside the bands. A bell curve, the textbook shape for random data, would put about 4.6% outside.
What are Bollinger Bands?
John Bollinger, a US market analyst, created the bands in the early 1980s while trading options. He set out the method in his 2001 book, Bollinger on Bollinger Bands.
The bands have three lines:
Middle band: the 20-day simple moving average, the mean of the last 20 closes. What Is a Moving Average? covers it.
Upper band: the middle band plus two standard deviations of those 20 closes.
Lower band: the middle band minus two standard deviations.
A standard deviation measures how far values typically sit from their mean. What Is Standard Deviation in Investing? works one out by hand. When prices swing more, the bands move apart. What Is Volatility? covers those swings.
How to calculate Bollinger Bands, step by step
The table lists SPY's 20 closes to 6 October 2026. SPY's $1.89 dividend came off its share price on 18 September. So the seven closes before that date are adjusted down by 0.25%, which keeps all 20 comparable.
Date (2026) | SPY close ($) | Close minus the mean ($) | Gap squared |
|---|---|---|---|
9 Sep | 760.51 | −4.55 | 20.69 |
10 Sep | 755.96 | −9.10 | 82.79 |
11 Sep | 762.40 | −2.66 | 7.07 |
14 Sep | 759.00 | −6.06 | 36.71 |
15 Sep | 755.52 | −9.54 | 90.99 |
16 Sep | 752.18 | −12.88 | 165.87 |
17 Sep | 760.71 | −4.35 | 18.91 |
18 Sep | 761.69 | −3.37 | 11.35 |
21 Sep | 773.50 | 8.44 | 71.25 |
22 Sep | 773.38 | 8.32 | 69.24 |
23 Sep | 767.81 | 2.75 | 7.57 |
24 Sep | 767.18 | 2.12 | 4.50 |
25 Sep | 771.35 | 6.29 | 39.58 |
28 Sep | 765.61 | 0.55 | 0.30 |
29 Sep | 764.20 | −0.86 | 0.74 |
30 Sep | 762.63 | −2.43 | 5.90 |
1 Oct | 763.99 | −1.07 | 1.14 |
2 Oct | 769.64 | 4.58 | 20.98 |
5 Oct | 774.83 | 9.77 | 95.47 |
6 Oct | 779.09 | 14.03 | 196.87 |
Mean | 765.06 | 0.00 | 47.40 |
Source: YX Insights
There are four steps:
Find the middle band. The 20 closes add up to $15,301.18. Divided by 20, the mean is $765.06.
Find the standard deviation. Take each close minus the mean. Square each gap, then take the mean of the squares: 47.40. Its square root is $6.88.
Draw the bands. Two standard deviations make $13.77. The upper band is $765.06 plus $13.77, or $778.83. The lower band is $765.06 minus $13.77, or $751.29.
Compare the close. SPY closed at $779.09, 26 cents above the upper band.
Step 2 divides by 20, as Bollinger's website does. Some guides divide by one less than the count. That gives slightly wider bands.
%b and BandWidth: where the price sits and how wide the bands are
Bollinger added two measures. %b shows where the close sits between the bands. It is the close minus the lower band, divided by the gap between the bands.
A %b of 0 is the lower band, while 1 is the upper band. On 6 October, SPY closed $27.80 above its lower band. The bands were $27.54 apart. So its %b was $27.80 divided by $27.54, or 1.01.
A %b above 1 means a close above the upper band.
BandWidth shows how wide the bands are. It is the gap between the bands divided by the middle band. On 6 October, SPY's BandWidth was $27.54 divided by $765.06, or 3.60%.
SPY's median BandWidth since 1993 is 5.37%. It peaked at 43.3% on 23 March 2020.
SPY's Bollinger Bands from January 2025 to October 2026
Source: YX Insights
Chart 1 shows SPY with its bands, with %b below. SPY closed outside the bands on 46 of 441 trading days. Of those, 25 were below the lower band and 21 above the upper band.
On 2 April 2025, the US announced a minimum 10% tariff on imports. SPY closed below its lower band the next day. On 4 April, its %b reached −0.40, the lowest in the chart. A month after 3 April, SPY was up 5.0%.
From 2 June to 30 September 2025, SPY closed above its upper band on 12 days, but never below its lower band. SPY rose 13.0% over those four months, with dividends counted.
How the Bollinger Band test works
The test counts each first close below the lower band. It does the same for each first close above the upper band. A first close means the close the day before was still inside that band.
Each signal uses the close of its own day. Its return runs to the close 21 trading days later, about one month. So no signal uses a later price.
The test covers 15 US-listed funds: the S&P 500, the Nasdaq-100, small companies, developed markets, nine sectors, gold and long Treasuries. Funds avoid survivorship bias, the error of testing only past winners, because no fund in the set was picked for doing well.
Each fund starts with its own history, from 1993 for SPY. Returns count dividends. Costs and tax are ignored.
Do Bollinger Bands work? Results on 15 funds
Source: YX Insights
Chart 2 compares the median one-month return after each signal with the median after all days, called the base rate.
For SPY, the median after a first close below the lower band was 2.21%, against 1.43% for all days. SPY rose in 70.7% of those months, against 65.7% for all days. The means were 1.58% and 0.96%.
After a first close above the upper band, SPY's median was 1.21%, a little below the base rate. SPY rose in 64.0% of those months, against 65.7% for all days.
Pooled across the 15 funds, the medians were 1.80% after lower-band closes, 1.15% after all days and 0.77% after upper-band closes. In all 15 funds, the lower-band median beat the base rate. In 14 of them, the upper-band median fell short of it.
The signals are not independent. Of SPY's 250 lower-band signals, 125 came within 21 trading days of the one before. Their months overlap.
To remove the overlap, a second count keeps a signal only if it came at least 21 trading days after the last signal kept. That leaves 134 signals, with a median of 1.90%.
Pooled signals also bunch in time. Of the pooled lower-band signals, 62.4% came in weeks when at least five funds triggered.
How often prices close outside Bollinger Bands
A bell curve is the symmetric shape where most values sit near the mean. Under a bell curve, 95.4% of values fall within two standard deviations, so about 4.6% fall outside.
Source: YX Insights
Chart 3 shows the share of daily closes outside the bands. SPY closed outside on 10.7% of days since 1993. Across the 15 funds, the share ran from 10.7% to 12.7%, with gold the highest. Pooled, it was 11.3%.
That matches Bollinger's own rules: "we typically find 90%, not 95%, of the data inside" the bands. He adds: "The distribution of security prices is non-normal." In other words, prices do not follow a bell curve.
What goes wrong with Bollinger Band signals
A lower-band close can come early in a fall. SPY first closed below its lower band on 25 February 2020. Over the next month, it fell 20.6%.
The outcomes spread wider. The worst 1/10 of months after a lower-band close lost more than 5.5%. For all days, the worst 1/10 lost more than 4.4%.
Bollinger warns against the textbook rule. His rules state: "Tags of the bands are just that, tags not signals."
How to read Bollinger Bands
Start with the defaults. These are 20 days and two standard deviations. For a 50-day average, Bollinger suggests 2.1 standard deviations.
Read %b and BandWidth together. %b says where the price sits. BandWidth says how wide the swings have been.
Pair the bands with another kind of gauge. Bollinger suggests a second indicator from a different family, such as volume or momentum. The Relative Strength Index (RSI) is a momentum gauge, covered in What Is RSI?
Bollinger Bands set a channel two standard deviations either side of a 20-day average, so it tracks volatility. On 15 funds, from 1993 or each fund's start, lower-band closes came before slightly better months than usual. Upper-band closes were mostly followed by gains.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
Common questions about Bollinger Bands
What does it mean when a price closes below the lower Bollinger Band?
It means the close sits more than two standard deviations below its 20-day average. John Bollinger says a touch of the band is not a buy signal on its own. For SPY since 1993, the median return in the month after a first close below the lower band was 2.21%, against 1.43% after all days.
What are the best settings for Bollinger Bands?
There is no single best setting. John Bollinger's defaults are a 20-day average with bands two standard deviations away. His rules suggest 2.1 standard deviations for a 50-day average and 1.9 for a ten-day one. He calls the defaults "just that, defaults". Each market may need its own.
How accurate are Bollinger Bands?
Bollinger Bands do not predict direction, so accuracy depends on the rule applied to them. Across 15 US-listed funds, prices rose in 60.4% of months after all days. After a first close below the lower band, they rose in 63.6% of months. After a first close above the upper band, they rose in 58.1%.
What percentage of prices stay inside Bollinger Bands?
About 88.7% of daily closes stayed inside the default bands across 15 US-listed funds, each from the start of its history. A bell curve would put 95.4% inside two standard deviations. John Bollinger's own rules say he typically finds "90%, not 95%" inside. He puts the gap down to prices not following a bell curve.
What is the difference between Bollinger Bands and RSI?
Bollinger Bands show where a price sits against its own recent swings. The Relative Strength Index (RSI) compares average gains with average losses over the last 14 days, on a scale of 0 to 100. Bollinger's rules suggest pairing the bands with an indicator of a different kind. What Is RSI? explains it.
DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.