Mean reversion bollinger

Bollinger Bands

Bullish: Price crosses above upper Bollinger Band. Bearish: Price crosses below lower Bollinger Band. Bands = SMA ± N standard deviations.

Signal family

Mean reversion — Oscillator-based signals that fire at overbought or oversold extremes — typically fade the prevailing move.

Parameters

Name Description Default Range
period SMA period 20 5–100
ub_factor Upper band std dev factor 2.0 0.5–4.0
lb_factor Lower band std dev factor 2.0 0.5–4.0

Historical context

382,833 valid triggers on 3,613 distinct tickers between 2015-03-10 and 2026-04-22. Universe: us_only · mcap ≥ $100,000,000 · price ≥ $1 (3,614 tickers). Entry at open T+1. 1d = intraday T+1; 20d = open T+1 to close T+20.

Benchmarks: spxew (S&P 500 Equal Weight — the primary benchmark here; a median-stock view that avoids the 2020+ megacap-concentration distortion), spx (S&P 500, cap-weighted), and msci (MSCI World USD). Per-stock regime: trending = ADX(14) ≥ 25, high vol = 20d ann. vol ≥ 20%.

At a glance (20d alpha vs S&P 500 Equal Weight, US-only)

Bullish
+0.12%
vs random-date null: worse than random (pperm=1.000)
Bearish (negative alpha = signal right)
+0.00%
vs random-date null: beats random (pperm=0.005)

Reading this: the random-date null is: for each ticker, sample N random dates and compute the same alpha — what alpha does a signal with no information produce? If the signal's observed alpha beats the null (pperm≤0.05), it's adding real information. If it's inside or worse than the null, the signal doesn't add value over random firing — any observed alpha is either noise or a universe artifact.

How often does BOLLINGER fire in each regime?

The signal's bucket distribution is itself informative. If 50%+ of all BOLLINGER triggers fire in the "non-trending + high vol" quadrant, the signal is structurally a chop-market event — regardless of what its textbook definition claims. Bullish and bearish are shown separately; counts are across the full US-only sample after the mcap and price floor.

Bollinger Bands (bollinger) — trigger count distribution by per-stock regime quadrant (trending/non-trending × high/low realized volatility) for , US-only universe

Per-stock regime quadrant — 20d alpha

Each trigger is tagged with the host stock's own technical regime on the trigger date: is the stock itself in a trend (ADX(14) ≥ 25) or ranging? And is its realized 20-day volatility high (≥ 20% annualized) or low? This is the textbook conditioning variable — "does this signal work better in trending stocks?" — answered at the level of the individual stock, not the market. Positive bars are good for the signal; negative bars mean alpha vanishes into the benchmark or worse.

Bollinger Bands (bollinger) — mean 20-day alpha versus S&P 500 Equal Weight by per-stock regime quadrant, side by side
Trending + Low vol
Stock in a clean directional move with low realized volatility. Textbook "trend-following paradise" — smooth grind with little whipsaw risk.
Trending + High vol
Violent directional moves — parabolic rallies, crisis selloffs. Trend exists but the path is noisy. Signal timing may be imprecise.
Non-trending + Low vol
Quiet chop, summer doldrums, consolidations. No directional bias but also no big swings — small edges become reliable if they exist at all.
Non-trending + High vol
Choppy and violent — the classical "whipsaw zone" for momentum signals. Crossovers and breakouts fire repeatedly without follow-through.

Sub-period check — does the signal work in every era?

A multi-year average can hide major instability. We split the sample into three non-overlapping windows: 2015–2019 (pre-COVID, normalized monetary policy), 2020–2022 (pandemic crash + recovery + rate-shock bear), and 2023+ (post-ZIRP, AI megacap rally). If a signal's alpha is positive overall but comes entirely from one era, that's a red flag — the conditions that produced it may not repeat. A robust signal shows a consistent sign across all non-empty buckets.

Bollinger Bands (bollinger) — 20-day alpha split by historical sub-period (2015-2019, 2020-2022, 2023+) to check consistency across market regimes

↑ Bullish triggers

Bench Metric 1d 5d 20d 60d 252d
spx Stock % -0.12% +0.10% +1.15% +3.69% +13.09%
Bench % -0.11% +0.15% +1.31% +3.82% +14.43%
Alpha % -0.04% -0.07% -0.16% -0.10% -1.29%
Median alpha -0.05% -0.19% -0.55% -1.60% -8.63%
Hit rate (α>0) 48.7% 47.9% 47.2% 45.7% 39.6%
p (naive) <0.001 <0.001 <0.001 0.0512 <0.001
p (HAC) <0.001 <0.001 <0.001 0.2275 0.0006
N 175,438 175,254 174,176 169,687 156,508
msci Stock % -0.12% +0.10% +1.15% +3.69% +13.09%
Bench % -0.04% +0.15% +1.17% +3.41% +12.34%
Alpha % -0.06% -0.09% -0.04% +0.29% +0.75%
Median alpha -0.11% -0.24% -0.46% -1.22% -6.48%
Hit rate (α>0) 47.4% 47.3% 47.8% 46.5% 42.0%
p (naive) <0.001 <0.001 0.1975 <0.001 <0.001
p (HAC) <0.001 <0.001 0.3003 0.0005 0.0493
N 174,978 173,527 172,814 168,795 155,966
spxew Stock % -0.12% +0.10% +1.15% +3.69% +13.09%
Bench % -0.14% +0.04% +1.04% +3.26% +11.10%
Alpha % -0.00% +0.05% +0.12% +0.47% +2.14%
Median alpha -0.06% -0.12% -0.30% -0.99% -5.28%
Hit rate (α>0) 48.6% 48.7% 48.5% 47.1% 43.2%
p (naive) 0.5598 0.0006 <0.001 <0.001 <0.001
p (HAC) 0.5636 0.0014 0.0004 <0.001 <0.001
N 174,680 173,789 172,328 168,244 154,604
Distribution of all 20d alpha outcomes for this direction. Median and winsorized mean shown.
Bollinger Bands (bollinger) — bullish 20-day alpha histogram showing distribution of per-trigger returns
Observed 20d alpha (vertical line) against the null distribution of random-date firing. If the line is deep inside the null cloud, the signal adds no information. If it sits in a tail, the signal is doing real work in that direction.
Bollinger Bands (bollinger) — bullish 20-day observed alpha versus random-date permutation null (200 iterations)
Permutation null detail — all horizons × both benchmarks
200-iteration null: for each ticker, sample N random dates from its history (matching observed trigger count) and compute the same alpha. The null distribution's 95% CI is where a signal with no information would land. pperm = one-sided fraction of null iters with mean ≥ observed.
Horizon Bench Observed α Null mean Null 95% CI pperm
1d spx -0.04% -0.00% [-0.03%, +0.46%] 0.990
1d msci -0.06% -0.02% [-0.05%, +0.44%] 0.995
1d spxew -0.00% -0.02% [-0.06%, +0.44%] 0.040
5d spx -0.07% +0.21% [-0.02%, +0.82%] 1.000
5d msci -0.09% +0.21% [-0.01%, +0.83%] 1.000
5d spxew +0.05% +0.23% [+0.00%, +0.85%] 0.781
20d spx -0.16% +0.30% [+0.00%, +0.99%] 1.000
20d msci -0.04% +0.42% [+0.13%, +1.12%] 1.000
20d spxew +0.12% +0.49% [+0.20%, +1.19%] 1.000
60d spx -0.10% +0.47% [+0.10%, +1.17%] 1.000
60d msci +0.29% +0.92% [+0.51%, +1.62%] 1.000
60d spxew +0.47% +1.14% [+0.76%, +1.86%] 1.000
252d spx -1.29% +0.66% [+0.10%, +1.63%] 1.000
252d msci +0.75% +2.99% [+2.43%, +3.94%] 1.000
252d spxew +2.14% +4.33% [+3.77%, +5.31%] 1.000

Example triggers on US large-caps (2023+, mcap ≥ $30B)

Six recent bullish BOLLINGER triggers on US mega-caps, filtered to |alpha| ≤ 25% to exclude catalyst-driven outliers (earnings surprises, M&A, binary events). The first three are the strongest outcomes — what the signal looks like when it works. The last three are the weakest — what the signal looks like when it fails. Each chart shows the stock's price with signal-appropriate technical overlays (e.g. MACD subpanel on MACD pages, Bollinger Bands on Bollinger pages, the 52-week trailing max line on 52w-high pages), a dot marking the trigger date, and the forward window shaded (green when the signal was right, red when it wasn't). Click any chart to open full-size.

Strongest outcomes (what BOLLINGER looks like when it works)
Bollinger Bands (bollinger) — strongest bullish trigger #1 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window Bollinger Bands (bollinger) — strongest bullish trigger #2 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window Bollinger Bands (bollinger) — strongest bullish trigger #3 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window
Weakest outcomes (what BOLLINGER looks like when it fails)
Bollinger Bands (bollinger) — weakest bullish trigger #1 on a US mega-cap, showing the 20-day forward window where the signal failed Bollinger Bands (bollinger) — weakest bullish trigger #2 on a US mega-cap, showing the 20-day forward window where the signal failed Bollinger Bands (bollinger) — weakest bullish trigger #3 on a US mega-cap, showing the 20-day forward window where the signal failed
Stock-regime quadrants (2×2 per-stock, 20d alpha detail table)
Each quadrant groups triggers by the stock's own ADX(14) and RV(20) at the trigger date — the textbook conditioning variable (not market-level). Stock %, bench %, alpha %, and HAC p-value shown for each benchmark.
Quadrant N Stock % (spx) Bench % (spx) Alpha % (spx) p (HAC) Stock % (msci) Bench % (msci) Alpha % (msci) p (HAC) Stock % (spxew) Bench % (spxew) Alpha % (spxew) p (HAC)
Trending + Low vol Clean directional grind, low whipsaw 12,254 +0.12% +0.80% -0.66% <0.001 +0.12% +0.62% -0.48% <0.001 +0.12% +0.57% -0.39% <0.001
Trending + High vol Crisis selloff or parabolic rally 63,391 +1.41% +1.63% -0.20% 0.0021 +1.41% +1.38% +0.04% 0.5386 +1.41% +1.12% +0.31% <0.001
Non-trending + Low vol Quiet chop, summer doldrums 17,996 +0.25% +0.82% -0.54% <0.001 +0.25% +0.66% -0.39% <0.001 +0.25% +0.52% -0.23% <0.001
Non-trending + High vol Classical "whipsaw zone" for momentum 81,865 +1.33% +1.31% +0.05% 0.3315 +1.33% +1.30% +0.08% 0.1417 +1.33% +1.24% +0.17% 0.0012
Sub-period breakdown table (20d alpha)
Historical clustering check. If alpha concentrates in one era, the signal's robustness is questionable.
Period N Alpha % (spx) p (HAC) Alpha % (msci) p (HAC) Alpha % (spxew) p (HAC)
2015-2019 2015-01-01 → 2020-01-01 50,784 +0.21% <0.001 +0.40% <0.001 +0.26% <0.001
2020-2022 2020-01-01 → 2023-01-01 58,336 -0.25% 0.0002 +0.04% 0.5366 -0.09% 0.1762
2023-2026 2023-01-01 → 2099-01-01 66,386 -0.37% <0.001 -0.43% <0.001 +0.24% <0.001

↓ Bearish triggers negative alpha = signal was right (stock underperformed market)

Bench Metric 1d 5d 20d 60d 252d
spx Stock % -0.05% +0.15% +0.64% +2.16% +11.98%
Bench % +0.01% +0.19% +0.93% +2.81% +13.31%
Alpha % -0.05% -0.03% -0.22% -0.62% -1.30%
Median alpha -0.05% -0.16% -0.79% -2.08% -8.13%
Hit rate (α>0) 48.5% 48.1% 45.8% 44.1% 39.9%
p (naive) <0.001 0.0195 <0.001 <0.001 <0.001
p (HAC) <0.001 0.0268 <0.001 <0.001 0.0004
N 207,233 206,062 204,083 200,578 178,456
msci Stock % -0.05% +0.15% +0.64% +2.16% +11.98%
Bench % +0.02% +0.17% +0.79% +2.34% +10.70%
Alpha % -0.07% -0.02% -0.07% -0.12% +1.27%
Median alpha -0.06% -0.15% -0.62% -1.58% -5.50%
Hit rate (α>0) 48.2% 48.2% 46.6% 45.4% 43.1%
p (naive) <0.001 0.1173 0.0044 0.0139 <0.001
p (HAC) <0.001 0.1373 0.0265 0.1499 0.0005
N 206,392 205,119 203,251 199,124 177,566
spxew Stock % -0.05% +0.15% +0.64% +2.16% +11.98%
Bench % +0.03% +0.19% +0.72% +1.99% +9.66%
Alpha % -0.07% -0.03% +0.00% +0.22% +2.52%
Median alpha -0.06% -0.14% -0.53% -1.24% -4.43%
Hit rate (α>0) 48.3% 48.4% 47.1% 46.4% 44.2%
p (naive) <0.001 0.0190 0.9168 <0.001 <0.001
p (HAC) <0.001 0.0265 0.9354 0.0053 <0.001
N 206,298 204,176 202,559 199,054 176,589
Distribution of all 20d alpha outcomes for this direction. Median and winsorized mean shown.
Bollinger Bands (bollinger) — bearish 20-day alpha histogram showing distribution of per-trigger returns
Observed 20d alpha (vertical line) against the null distribution of random-date firing. If the line is deep inside the null cloud, the signal adds no information. If it sits in a tail, the signal is doing real work in that direction.
Bollinger Bands (bollinger) — bearish 20-day observed alpha versus random-date permutation null (200 iterations)
Permutation null detail — all horizons × both benchmarks
200-iteration null: for each ticker, sample N random dates from its history (matching observed trigger count) and compute the same alpha. The null distribution's 95% CI is where a signal with no information would land. pperm = one-sided fraction of null iters with mean ≥ observed.
Horizon Bench Observed α Null mean Null 95% CI pperm
1d spx -0.05% -0.01% [-0.03%, -0.00%] 0.005
1d msci -0.07% -0.03% [-0.05%, -0.02%] 0.005
1d spxew -0.07% -0.03% [-0.05%, -0.02%] 0.005
5d spx -0.03% +0.20% [+0.01%, +0.80%] 0.005
5d msci -0.02% +0.21% [+0.01%, +0.81%] 0.005
5d spxew -0.03% +0.23% [+0.04%, +0.83%] 0.005
20d spx -0.22% +0.35% [+0.12%, +0.77%] 0.005
20d msci -0.07% +0.47% [+0.25%, +0.88%] 0.005
20d spxew +0.00% +0.54% [+0.32%, +0.96%] 0.005
60d spx -0.62% +0.74% [+0.49%, +1.17%] 0.005
60d msci -0.12% +1.19% [+0.93%, +1.60%] 0.005
60d spxew +0.22% +1.41% [+1.16%, +1.82%] 0.005
252d spx -1.30% +2.11% [+1.68%, +2.58%] 0.005
252d msci +1.27% +4.42% [+3.99%, +4.85%] 0.005
252d spxew +2.52% +5.78% [+5.34%, +6.27%] 0.005

Example triggers on US large-caps (2023+, mcap ≥ $30B)

Six recent bearish BOLLINGER triggers on US mega-caps, filtered to |alpha| ≤ 25% to exclude catalyst-driven outliers (earnings surprises, M&A, binary events). The first three are the strongest outcomes — what the signal looks like when it works. The last three are the weakest — what the signal looks like when it fails. Each chart shows the stock's price with signal-appropriate technical overlays (e.g. MACD subpanel on MACD pages, Bollinger Bands on Bollinger pages, the 52-week trailing max line on 52w-high pages), a dot marking the trigger date, and the forward window shaded (green when the signal was right, red when it wasn't). Click any chart to open full-size.

Strongest outcomes (what BOLLINGER looks like when it works)
Bollinger Bands (bollinger) — strongest bearish trigger #1 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window Bollinger Bands (bollinger) — strongest bearish trigger #2 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window Bollinger Bands (bollinger) — strongest bearish trigger #3 on a US mega-cap, showing the stock's OHLC candles, moving averages, signal-specific overlays, and the 20-day forward window
Weakest outcomes (what BOLLINGER looks like when it fails)
Bollinger Bands (bollinger) — weakest bearish trigger #1 on a US mega-cap, showing the 20-day forward window where the signal failed Bollinger Bands (bollinger) — weakest bearish trigger #2 on a US mega-cap, showing the 20-day forward window where the signal failed Bollinger Bands (bollinger) — weakest bearish trigger #3 on a US mega-cap, showing the 20-day forward window where the signal failed
Stock-regime quadrants (2×2 per-stock, 20d alpha detail table)
Each quadrant groups triggers by the stock's own ADX(14) and RV(20) at the trigger date — the textbook conditioning variable (not market-level). Stock %, bench %, alpha %, and HAC p-value shown for each benchmark.
Quadrant N Stock % (spx) Bench % (spx) Alpha % (spx) p (HAC) Stock % (msci) Bench % (msci) Alpha % (msci) p (HAC) Stock % (spxew) Bench % (spxew) Alpha % (spxew) p (HAC)
Trending + Low vol Clean directional grind, low whipsaw 21,083 +0.37% +0.46% -0.07% 0.1595 +0.37% +0.25% +0.13% 0.0119 +0.37% +0.11% +0.27% <0.001
Trending + High vol Crisis selloff or parabolic rally 72,644 +0.82% +1.03% -0.15% 0.0325 +0.82% +0.90% +0.01% 0.9448 +0.82% +0.83% +0.09% 0.2079
Non-trending + Low vol Quiet chop, summer doldrums 24,031 +0.28% +0.62% -0.34% <0.001 +0.28% +0.41% -0.11% 0.0144 +0.28% +0.32% -0.02% 0.6870
Non-trending + High vol Classical "whipsaw zone" for momentum 89,563 +0.78% +1.03% -0.20% <0.001 +0.78% +0.91% -0.08% 0.1015 +0.78% +0.86% -0.03% 0.5324
Sub-period breakdown table (20d alpha)
Historical clustering check. If alpha concentrates in one era, the signal's robustness is questionable.
Period N Alpha % (spx) p (HAC) Alpha % (msci) p (HAC) Alpha % (spxew) p (HAC)
2015-2019 2015-01-01 → 2020-01-01 63,026 -0.49% <0.001 -0.26% <0.001 -0.18% <0.001
2020-2022 2020-01-01 → 2023-01-01 63,408 +0.02% 0.7681 +0.14% 0.0282 -0.12% 0.0711
2023-2026 2023-01-01 → 2099-01-01 80,893 -0.17% 0.0075 -0.05% 0.3887 +0.28% <0.001

Methodology and caveats

How to read. Entry at open of T+1 (one trading day after the signal fires on close of T). 20d = open T+1 to close T+20. Alpha = stock return − benchmark return over the same window (Convention A, single-sided, textbook). For bullish triggers, POSITIVE alpha = signal was right. For bearish triggers, NEGATIVE alpha = signal was right (stock underperformed market). No sign-flipping; the direction of the bet determines what "good" looks like. Per-stock regime is each stock's own ADX(14) and RV(20) at the trigger date — not market-wide state.

Three p-values, three robustness tests. (a) p_naive: scipy one-sample t-test on winsorized alphas. Optimistic because overlapping 20d windows on the same ticker inflate effective N. (b) p_hac: Newey-West HAC with lag = horizon — corrects for the overlap and is the academic-finance standard. (c) p_perm: fraction of 200 random-date null iterations with mean ≥ observed. Tests whether the signal beats random date selection at all. A signal that clears all three (pnaive, phac, pperm all < 0.05) has real information; a signal that fails pperm has zero edge even if the t-test says "significant."

Caveats. (i) Universe reflects today's active tickers; delisted losers pruned → survivorship bias. (ii) Mcap ≥ $100M filter uses today's snapshot, not point-in-time — mild lookahead on which stocks enter the sample, not on returns. (iii) Means and p-values use winsorized alphas (1/99 percentile) to prevent data errors from dominating. Medians and hit rates use raw data. (iv) Zero transaction costs assumed. Realistic bid-ask + commissions remove 20–40bps from 20d alpha on US large-caps, more on small-cap. Sub-20bps alpha is noise in practice. (v) Past performance does not predict future results.

How to use this

1 · When to reach for this signal

Caution recommended. Bullish 20d alpha is -0.16% and worse than random — triggering on random dates would have produced better long-side returns. Either direction fails the "beats random" test. Don't use Bollinger Bands as a standalone entry trigger. It may still be useful as part of a composite (section 4).

2 · When it works — the setups that drive it

  • Best bullish setup: Non-trending + High vol — alpha +0.05% / 20d on 81,865 historical triggers.
  • Best bearish setup: Trending + Low vol — alpha -0.07% / 20d on 21,083 historical triggers.
  • Best era for bullish: 2015-2019 — alpha +0.21% / 20d.
  • Best era for bearish: 2020-2022 — alpha +0.02% / 20d.

3 · When it fails — common false positives

  • Weakest bullish cell: Trending + Low vol — alpha -0.66% / 20d on 12,254 triggers.
  • Weakest bearish cell: Non-trending + Low vol — alpha -0.34% / 20d on 24,031 triggers.
  • Worst era for bullish: 2023-2026 — alpha -0.37% / 20d.
  • Worst era for bearish: 2015-2019 — alpha -0.49% / 20d.

Signal-specific failure patterns

Bullish Bollinger fails against every benchmark, every horizon
Lower-band touch bullish signal produces negative alpha at 1d, 5d, 20d and 60d across all four benchmarks (p_perm=1.000 vs each). The band-touch thesis — 'price has stretched too far from its mean, expect reversion' — does not survive outside of a bounded range-bound stock. In a trending universe (US large-caps 2015-2026), stocks touching the lower band are usually doing so because they have real bad news, not because they've randomly wandered below a statistical envelope.
evidence: bullish 20d vs SPX α=−0.16 p_perm=1.000; 60d α=−0.10 p_perm=1.000
Bullish worst in trending regimes, marginal in high-vol chop
Within the bullish side, the least damaging regime is non-trend high-vol at α=+0.05 (essentially zero). The deepest loss is trending_low_vol at α=−0.66. Translation: lower-band touches in clean uptrends are catching falling stocks — the trend just broke. Don't dip-buy Bollinger lower touches on winners that have started to crack.
evidence: 20d bullish by regime vs SPX: trending_low_vol −0.66 (worst), nontrend_high_vol +0.05 (least bad)
Bearish side has consistent, compounding short-side edge
Upper-band touch bearish delivers α=−0.22 at 20d (p(HAC)<1e-11, p_perm=0.005) and widens to −0.62 at 60d (p<1e-14, p_perm=0.005). The effect is robust across sub-periods (2015-2019 −0.49, 2023-2026 −0.17; only 2020-2022 is noisy at +0.02). Upper-band stretched stocks tend to mean-revert while the broader market continues up.
evidence: bearish vs SPX: 20d α=−0.22 p_perm=0.005; 60d α=−0.62 p_perm=0.005
Pre-COVID 2015-2019 bullish was positive — bullish broke post-pandemic
Sub-period breakdown shows bullish α=+0.22 in 2015-2019 (small but positive), then −0.25 in 2020-2022 and −0.37 in 2023-2026. Like RSI, Bollinger bullish is a pre-QE artifact. Whatever mean-reversion dynamic worked in the low-rate low-growth 2015-2019 market evaporated when stimulus-driven capital concentration took over.
evidence: bullish 20d vs SPX: 2015-2019 +0.22, 2020-2022 −0.25, 2023-2026 −0.37

4 · Pairing inside a screen

The statements below describe how this signal relates to others by construction — which indicator family it belongs to, and where same-family redundancy might reduce the independence of evidence inside a Daily Report. These are taxonomic classifications drawn from standard technical-analysis texts; they are not pairing backtests. A multi-signal convergence backtest is planned but not yet run.

Volatility-envelope construction

Bollinger Bands are a volatility envelope: a 20-period simple moving average ± 2 standard deviations of price (Bollinger, Bollinger on Bollinger Bands, 2001). This construction is distinct from momentum oscillators (RSI, Stochastics, Williams %R, CCI) and from moving-average crossover signals, so pairing Bollinger with any one of them does not produce same-family redundancy.

What would likely rescue this signal

This block calls out the data or conditions that could turn a technically weak signal into a usable one in a composite screen. Based on signal mechanics and the observed failure patterns above; individual combinations are not yet backtested.

  • Volume-filter the bullish side — A lower-band touch on HIGH volume is capitulation; on low volume is drift. A bullish filter requiring volume > 1.5× average + a follow-through close above T+1 open might isolate real reversals from continuing drawdowns. Not currently implemented in the screen filter layer.
  • Regime gate the bullish side — Bullish Bollinger worked 2015-2019 and broke 2020+. A filter requiring 'market breadth > 55%' (most stocks above 50DMA) or 'SPX not within 5% of ATH' might carve out the residual mean-reversion regime.
  • Hold the bearish trade to 60d — Bearish edge doubles from 20d to 60d (−0.22 → −0.62). Short-horizon exits on Bollinger bearish leave alpha on the table. This argues for a time-stop rather than a profit-target exit — hold the full 60d window absent a structural invalidation (stock taking out its upper-band high by > 2%).

See also Why technical-only signals don't survive on their own for the broader argument.

5 · Before you act — a 5-point checklist

  1. Normal trading day? Rule out earnings (within ±3 days), ex-dividend, or known corporate-action dates — the signal is almost certainly reading noise, not momentum, in those windows.
  2. Where is price vs its own 50 / 200 DMA? A mean-reversion signal firing against the long-term trend (e.g. oversold in a clean uptrend) is much more reliable than one firing with it.
  3. What's the sector breadth doing? An isolated signal in a broadly down-trending sector is a lower-confidence setup than one firing with the rest of its peer group.
  4. Is ADV20 enough for your size? If the trigger is on a $500M name and you want to move $1M notional, you're the tape. Consider adv20d ≥ 5% of your intended position.
  5. What invalidates you? Define a price level (for longs: a close below the trigger-day low; for shorts: close above the trigger-day high) and honor it. The backtest alpha is an average; any one trade can be at either tail.

Execution notes

Bearish upper-band touches are the only tradable direction here. 20d and 60d both deliver real alpha; 60d compounds more but adds event risk. Entry = open T+1. The bullish lower-touch trade is a structural loser — stocks touching their lower band in a bull market are usually breaking down, not pausing. Note: Bollinger Bands period/std-dev parameters matter a lot; this backtest uses 20-period / 2 stdev which is the Lambert default. Tighter parameters (shorter window, lower stdev) would fire more often on smaller deviations — likely noisier.