Does a Bollinger band breakout lead to more gains?
The opposite, by a small margin. Since 1995, US stocks closing above their upper Bollinger band beat the same day's median stock over the next week 49 times in 100; a coin flip gives 50. The shortfall is proven, lagging in 25 of 32 years, and gone within a month (49). Closes below the lower band beat it 51 times in 100: faint, not reliable on its own.
The paper
The full study: its data, method, robustness, limits and sources.
Opulence Alpha Research · Published Sep 25, 2026 · Data through Aug 21, 2026
Keywords: Bollinger bands, Bollinger band breakout, do Bollinger bands work, upper Bollinger band, lower Bollinger band, Bollinger band strategy backtest·JEL classification: G11, G12, G14, C12, C58
- Universe
- The same for every study: 1,767 US common stocks in the 11 GICS sectors, 422 of them since delisted; S&P 500 members since 1996 plus large and mid-sized companies outside the index
- Period
- Every Wednesday from Jan 4, 1995 to Aug 19, 2026: 1,634 Wednesdays across 7,959 trading sessions
- Sample
- n = 1,763 stocks of the 1,767-stock universe; 1,889,758 stock-weeks, a median of 1,148 stocks per Wednesday
- Outcome
- Return over the next 1, 5, 21 and 63 trading sessions against the same day's median stock
- Inference
- t on non-overlapping dates; proven only when |t| ≥ 3 and the same sign in at least 60% of years
1Introduction
John Bollinger's bands sit two standard deviations above and below a 20-day moving average, so a close outside them marks an unusually large move for that stock. Two readings circulate: that the bands are limits the price tends to return inside, and, in the version tested here, that a close above the upper band is a breakout, a show of strength that tends to continue, with a close below the lower band as its bearish mirror. This study asks whether stocks closing above the upper band go on to beat other stocks, and whether stocks closing below the lower band go on to trail them.
2Data and method
2.1Sample design
Population and frame. The population is US common stocks listed on the NYSE and Nasdaq; funds, ETFs, trusts, preferred shares, warrants and units are excluded. The sampling frame is a fixed universe of 1,767 companies, drawn once when the platform was built and not re-sampled since, in two strata. Stratum 1 is a census of the S&P 500: every company in the index at any time since 1996 whose price history could be recovered, 1,027 companies of which 413 have since delisted; it holds 76% of the index's members in 1996 and at least 96% in every year from 2010. Stratum 2 is 740 large and mid-sized companies outside the index, selected in proportion to the market's sector weights from the stocks that passed a minimum share price of $15 and a minimum average daily trading value of $25 million; 9 of them have since delisted.
Sample. The unit of observation is a stock-week: one stock on one Wednesday. A stock enters a Wednesday's cross-section when it has a valid close that day, a value of the signal and a measured outcome; bars flagged as bad data and returns that cross a change of issuer are left out. This study's sample is n = 1,763 stocks of the 1,767: 1,889,758 stock-weeks on 1,634 Wednesdays from Jan 4, 1995 to Aug 19, 2026, a median of 1,148 stocks per Wednesday (range 820 to 1,371). 331,954 stock-weeks (17.6%) come from the 422 companies that have since delisted. Every stock and every Wednesday carries equal weight.
Representativeness. Table 1 gives the sample by GICS sector beside the S&P Composite 1500: 7.9% of companies would have to change sector for the two to match exactly. By latest market value, 55% of the active companies are large (at least $10bn), 38% mid ($2–10bn) and 6% small. Because Stratum 2 was chosen from companies listed at construction, its history carries survivorship bias. Section 4 reads the result horizon by horizon and year by year.
Companies in this study's sample, of them those since delisted, the size of the active companies, the sample's share of stock-weeks and the sector's share of the S&P Composite 1500.
| Sector | Sample | Delisted | Large | Mid | Small | Stock-weeks | S&P 1500 |
|---|---|---|---|---|---|---|---|
| Information Technology | 282 | 67 | 122 | 76 | 18 | 13.9% | 12.7% |
| Financials | 256 | 60 | 114 | 74 | 8 | 14.6% | 17.2% |
| Industrials | 255 | 53 | 126 | 67 | 11 | 15.9% | 17.5% |
| Health Care | 237 | 47 | 92 | 81 | 16 | 12.4% | 10.9% |
| Consumer Discretionary | 209 | 39 | 69 | 87 | 12 | 12.3% | 12.9% |
| Energy | 105 | 30 | 39 | 31 | 5 | 5.9% | 4.7% |
| Consumer Staples | 101 | 38 | 40 | 17 | 5 | 5.9% | 4.9% |
| Materials | 94 | 34 | 35 | 24 | 2 | 5.6% | 5.1% |
| Real Estate | 84 | 10 | 40 | 32 | 2 | 5.7% | 6.9% |
| Communication Services | 80 | 31 | 28 | 13 | 5 | 3.9% | 3.3% |
| Utilities | 60 | 13 | 35 | 12 | 0 | 4% | 4% |
| All sectors | 1,763 | 422 | 740 | 514 | 84 | 100% | 100% |
Sample: n = 1,763 of the 1,767 companies. Size by latest market value for the 1,338 active companies with one: large ≥ $10bn, mid $2–10bn, small < $2bn. S&P 1500 shares count the constituents of the S&P 500, MidCap 400 and SmallCap 600 (1,506 companies, lists read Sep 25, 2026). Sectors are each company's current GICS sector.
2.2Signals
| Signal | Definition | Since | Wednesdays |
|---|---|---|---|
| Close above the upper band | The close sits above the 20-day average plus 2 standard deviations, on any observed day | Jan 4, 1995 | 1,630 |
| Close below the lower band | The close sits below the 20-day average minus 2 standard deviations, on any observed day | Jan 4, 1995 | 1,626 |
2.3Measurement
Outcome. The return from the close on the Wednesday to the close 1, 5, 21 and 63 trading sessions later, on closes adjusted for splits and dividends, compared with the same day's median stock; half of all stocks beat the median by construction, so chance is 50 in 100 on every date. No delisting return is added.
Each Wednesday, the stocks showing the reading are scored against the same day's median stock. A state counts on every Wednesday it holds; a cross or breakout counts only on the Wednesday it happens.
The t-statistic uses non-overlapping dates only. A result is called proven when |t| is at least 3, it held in at least 60% of years and it is large enough to matter; with 1,470 tests across the studies, a looser bar would pass dozens by luck.
Full data and methods3Results
Figure 1 reads every horizon for both bands. Stocks closing above the upper band beat the median stock 49 times in 100 over the next session, 49 over the next week, 49 over the next month and 49 over the next three months. The session and the week are proven relations against the breakout; the month and the three months show no measurable edge. Stocks closing below the lower band, the breakdown, beat the median stock 51, 51, 51 and 51 times in 100 over the same horizons: faint at the two short horizons, not reliable on its own, and no measurable edge beyond them.
Share of stocks showing the reading that beat the same day's median stock, by horizon. The line at 50 is chance, the grey band the range chance alone produces; a filled square is a proven relation.
Each square is one calendar year; filled = a year in which the average stock showing the signal lagged the median stock over the next week. Years are counted, not shown in order.
A signal with an edge would fill most squares, or leave most empty.
Out of every 100 stocks above the upper band, 49 beat the median stock over the following week; out of every 100 below the lower band, 51 did (Figure 2). The breakout trailed, a proven relation; the breakdown's slight lead is faint and not reliable on its own. Both are tendencies across 104K and 78.1K stock-weeks, not forecasts for any one stock.
4Robustness
Table 3 reads each band horizon by horizon. Over the next week, stocks above the upper band lagged the median stock in 25 of 32 years, and stocks below the lower band beat it in 23 of 32, a faint lead. Over the next month neither reading clears the bar for an edge (t -1.8 and 0.9).
Share that beat the median stock, its t-statistic on non-overlapping dates, and the years in which it pointed the same way.
| Horizon | Beat the median | t | Years, same way |
|---|---|---|---|
| Close above the upper band, session | 48.6 | -5.2 | 24 / 32 |
| Close above the upper band, week | 48.6 | -5.6 | 25 / 32 |
| Close above the upper band, month | 49.1 | -1.8 | 21 / 32 |
| Close above the upper band, three months | 49.1 | -1.4 | 21 / 32 |
| Close below the lower band, session | 50.8 | 2.4 | 21 / 32 |
| Close below the lower band, week | 50.9 | 2.5 | 23 / 32 |
| Close below the lower band, month | 50.6 | 0.9 | 19 / 32 |
| Close below the lower band, three months | 50.8 | 0.4 | 23 / 32 |
5Limitations
- The reading is a state, not a first touch: a stock is counted on every Wednesday its close sits outside a band, so a stock that rides the upper band for weeks appears more than once. Only the textbook setting, 20 days and 2 standard deviations, is measured.
- Before costs. Averages exclude trading costs, taxes and market impact.
- Same-close timing. Returns start at the close the signal is computed from; a real trade would start later.
- Survivors among smaller companies. The non-index names were chosen from companies listed when the universe was built; no delisting returns are added.
6Conclusion
A close outside the Bollinger bands measures a move that has already been made. Stocks that closed above the upper band did not go on to beat other stocks: over the following week they trailed them, by a small margin that held in most years, and after a month they were indistinguishable from the rest. Stocks below the lower band did not go on to trail. Read as a breakout, the band's signal runs the wrong way; whichever way it is read, what little there is fades within a month.
References
- Bollinger, J. (2001). Bollinger on Bollinger Bands. McGraw-Hill.
- Jegadeesh, N. (1990). Evidence of Predictable Behavior of Security Returns. Journal of Finance 45(3).
- Park, C.-H. and Irwin, S. H. (2007). What Do We Know About the Profitability of Technical Analysis? Journal of Economic Surveys 21(4).
- Benjamini, Y. and Hochberg, Y. (1995). Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society, Series B 57(1).
- Grinold, R. C. and Kahn, R. N. (2000). Active Portfolio Management, 2nd ed. McGraw-Hill.
- Harvey, C. R., Liu, Y. and Zhu, H. (2016). … and the Cross-Section of Expected Returns. Review of Financial Studies 29(1).
- Shumway, T. (1997). The Delisting Bias in CRSP Data. Journal of Finance 52(1).
Appendix A. Questions readers ask
- Is a close above the upper Bollinger band bullish?
- Not in the data. Since 1995, US stocks closing above the upper band beat the same day's median stock over the next week 49 times in 100, against 50 for a coin flip. That small shortfall is a proven relation, and it runs against the breakout reading; over the next month no measurable edge remained (49 in 100).
- Does a close below the lower band mean more losses?
- Not reliably. Stocks closing below the lower band beat the median stock over the next week 51 times in 100, a little above a coin flip: faint, not reliable on its own, and with no measurable edge over the next month (51).
- Does the upper band act as a ceiling instead?
- Only a little, and only briefly. Stocks above the band lagged the median stock over the next week by a small margin, in line with the short-term reversal long documented in US stocks (Jegadeesh 1990), and over the next month no measurable edge remained. John Bollinger himself warned against reading a touch of a band as a signal on its own.
- How were the Bollinger bands measured?
- Every Wednesday from Jan 4, 1995 to Aug 19, 2026, the stocks whose close sat above the upper band, or below the lower band, of a 20-day average plus or minus 2 standard deviations were compared with the same day's median stock over the next 1, 5, 21 and 63 trading sessions, on the same universe of 1,767 US stocks every study uses.
Data availability and citation
Every figure in this paper is quoted from one frozen snapshot, published as JSON with the sample description, the robustness results and the test counts. The same snapshot feeds the research console, so the two cannot disagree.
Opulence Alpha Research (2026). Does a Bollinger band breakout lead to more gains? Opulence Alpha Studies, Sep 25, 2026. https://opulencealpha.ai/studies/bollinger-band-breakout
Research, not advice. A measured tendency across hundreds of stocks is a nudge for any one of them, never a forecast of its price.