new_high_low
20-Day New High / New Low
Detects when price makes a new rolling high or low. Bullish: today's high = highest high over N days AND close >= open. Bearish: today's low = lowest low over N days AND close <= open. The candle-direction filter removes false signals from intraday wicks.
Signal family
Trend — Signals that fire when price is continuing or reversing an established directional move. Momentum-following by nature.
Parameters
| Name | Description | Default | Range |
|---|---|---|---|
| period | Lookback period (days) | 20 | 5–252 |
Historical context
7,744,906 triggers on 24,073 tickers, 1988-03-14 → 2026-05-01. Universe: global — all covered exchanges (mcap ≥ $100,000,000, price ≥ $1). Long-only convention: BUY at open T+1, hold the horizon, compare to S&P 500 Equal Weight over the same window.
Methodology footnotes
Benchmarks shown in the detail tables: spxew (S&P 500 Equal Weight — primary, median-stock view, avoids the 2020+ megacap-concentration distortion), spx (S&P 500 cap-weighted, distorted post-2020), msci (MSCI World USD). Per-stock regime tags: trending = ADX(14) ≥ 25, high vol = 20d realized annualized vol ≥ 20%. 1d return = intraday T+1 open→close; 20d = open T+1 to close T+20.
At a glance — alpha vs S&P 500 Equal Weight, global universe
Holding-period sensitivity. Bullish columns: positive = signal worked (long the trigger beat the index). Bearish columns: negative = signal worked (the flagged stock underperformed).
| Horizon | Bullish α | Bearish α |
|---|---|---|
| 5-day | -0.06% | +0.03% |
| 20-day | +0.13% | +0.08% |
| 60-day | +0.41% | +0.20% |
| 1-year | +3.08% | +1.80% |
Bearish: worse than random (p=1.000).
Where does NEW_HIGH_LOW actually fire?
The bucket distribution often reveals what the signal really is, regardless of its textbook label. Heavy concentration in "non-trending + high vol" = it's mostly a chop-market event. Heavy in "trending + low vol" = it picks up the smooth grinds. Read the chart before the alpha numbers — context shapes everything that follows.
Does it work in every regime?
Trigger alpha split by the host stock's own regime on the trigger date — trending or ranging, high-vol or low-vol. This is the 20-day alpha a trade taken on the trigger would have captured. Bars carrying the sign that favours the trade (positive for bullish triggers, negative for bearish) mark the regimes where the signal worked; the opposite sign marks regimes to avoid. One strong bar beside three flat ones is not a "20-day alpha" signal — it is a "20-day alpha when the stock is X" signal. Bar labels carry the sample size; a cell built on a handful of triggers is noise, not a regime finding.
Does it work in every era?
A multi-year average can hide major instability. The sample splits into three windows: 2015–2019 (pre-COVID), 2020–2022 (pandemic + 2022 bear), and 2023+ (post-ZIRP + AI megacap rally). All three carrying the sign that favours the trade means the signal is durable; one era doing all the work means a regime-specific edge that may not repeat. The greater the variance across eras, the smaller the position it justifies.
↑ Bullish triggers
| Bench | Metric | 1d | 5d | 20d | 60d | 252d |
|---|---|---|---|---|---|---|
| spxew | Stock % | -0.00% | +0.12% | +0.79% | +2.50% | +12.85% |
| Bench % | +0.04% | +0.18% | +0.69% | +2.11% | +9.88% | |
| Alpha % | -0.04% | -0.06% | +0.13% | +0.41% | +3.08% | |
| Median alpha | -0.09% | -0.31% | -0.81% | -1.98% | -6.20% | |
| Hit rate (α>0) | 47.2% | 46.3% | 45.5% | 44.1% | 41.9% | |
| p (naive) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| p (HAC) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| N | 3,932,787 | 3,775,570 | 3,757,841 | 3,692,389 | 3,249,414 | |
| spx | Stock % | -0.00% | +0.12% | +0.79% | +2.50% | +12.85% |
| Bench % | +0.02% | +0.22% | +0.89% | +2.95% | +13.63% | |
| Alpha % | -0.03% | -0.09% | -0.08% | -0.44% | -0.83% | |
| Median alpha | -0.08% | -0.35% | -1.09% | -2.89% | -10.42% | |
| Hit rate (α>0) | 47.2% | 45.6% | 44.1% | 41.6% | 37.4% | |
| p (naive) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| p (HAC) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| N | 3,967,057 | 3,832,673 | 3,800,120 | 3,741,473 | 3,294,451 | |
| msci | Stock % | -0.00% | +0.12% | +0.79% | +2.50% | +12.85% |
| Bench % | +0.05% | +0.20% | +0.78% | +2.50% | +11.13% | |
| Alpha % | -0.05% | -0.08% | +0.05% | +0.02% | +1.63% | |
| Median alpha | -0.11% | -0.34% | -0.96% | -2.42% | -7.98% | |
| Hit rate (α>0) | 46.5% | 45.7% | 44.7% | 42.7% | 40.2% | |
| p (naive) | <0.001 | <0.001 | <0.001 | 0.1252 | <0.001 | |
| p (HAC) | <0.001 | <0.001 | 0.0002 | 0.6532 | <0.001 | |
| N | 3,940,348 | 3,802,594 | 3,789,418 | 3,713,092 | 3,273,623 |
Permutation null detail — all horizons × each benchmark
| Horizon | Bench | Observed lift | Null mean | Null 95% CI | pperm |
|---|---|---|---|---|---|
| 1d | spxew | +0.07% | +0.07% | [+0.07%, +0.08%] | 0.239 |
| 1d | spx | +0.08% | +0.08% | [+0.08%, +0.08%] | 0.995 |
| 1d | msci | +0.08% | +0.09% | [+0.08%, +0.09%] | 1.000 |
| 5d | spxew | +0.28% | +0.32% | [+0.32%, +0.33%] | 1.000 |
| 5d | spx | +0.30% | +0.34% | [+0.34%, +0.35%] | 1.000 |
| 5d | msci | +0.30% | +0.35% | [+0.34%, +0.35%] | 1.000 |
| 20d | spxew | +1.06% | +1.06% | [+1.05%, +1.07%] | 0.507 |
| 20d | spx | +1.09% | +1.10% | [+1.09%, +1.11%] | 0.851 |
| 20d | msci | +1.10% | +1.11% | [+1.10%, +1.12%] | 0.881 |
| 60d | spxew | +2.00% | +2.26% | [+2.24%, +2.28%] | 1.000 |
| 60d | spx | +1.91% | +2.35% | [+2.32%, +2.37%] | 1.000 |
| 60d | msci | +1.94% | +2.36% | [+2.33%, +2.37%] | 1.000 |
| 252d | spxew | +2.98% | +3.87% | [+3.83%, +3.92%] | 1.000 |
| 252d | spx | +3.25% | +4.31% | [+4.27%, +4.36%] | 1.000 |
| 252d | msci | +3.33% | +4.24% | [+4.20%, +4.30%] | 1.000 |
Example triggers on US large-caps (2023+, mcap ≥ $30B)
Six recent bullish NEW_HIGH_LOW triggers on US mega-caps, ranked within this sample of six: the top three and the bottom three, with extreme outliers excluded. Ranking is relative, not absolute — where the signal did well across the sampled names, even the bottom three can have beaten the benchmark, so the alpha printed on each panel is what settles it. Both groups are tail outcomes by construction; read them as the range, not the typical result.
Best three of the six sampled
Weakest three of the six sampled — not necessarily losses
Stock-regime quadrants (2×2 per-stock, 20d alpha detail table)
| Quadrant | N | Stock % (spxew) | Bench % (spxew) | Alpha % (spxew) | p (HAC) | Stock % (spx) | Bench % (spx) | Alpha % (spx) | p (HAC) | Stock % (msci) | Bench % (msci) | Alpha % (msci) | p (HAC) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Trending + Low vol Clean directional grind, low whipsaw | 616,012 | +0.51% | +0.55% | +0.04% | 0.2442 | +0.51% | +0.79% | -0.23% | <0.001 | +0.51% | +0.67% | -0.09% | 0.0017 |
| Trending + High vol Crisis selloff or parabolic rally | 1,397,458 | +0.95% | +0.69% | +0.30% | <0.001 | +0.95% | +0.92% | +0.06% | 0.0186 | +0.95% | +0.80% | +0.20% | <0.001 |
| Non-trending + Low vol Quiet chop, summer doldrums | 601,562 | +0.48% | +0.50% | +0.01% | 0.6576 | +0.48% | +0.74% | -0.24% | <0.001 | +0.48% | +0.61% | -0.10% | <0.001 |
| Non-trending + High vol Classical "whipsaw zone" for momentum | 1,429,939 | +0.94% | +0.82% | +0.14% | <0.001 | +0.94% | +0.96% | -0.01% | 0.5698 | +0.94% | +0.87% | +0.09% | <0.001 |
Sub-period breakdown table (20d alpha)
| Period | N | Alpha % (spxew) | p (HAC) | Alpha % (spx) | p (HAC) | Alpha % (msci) | p (HAC) |
|---|---|---|---|---|---|---|---|
| 2015-2019 2015-01-01 → 2020-01-01 | 1,357,451 | -0.37% | <0.001 | -0.60% | <0.001 | -0.41% | <0.001 |
| 2020-2022 2020-01-01 → 2023-01-01 | 1,153,081 | +0.09% | 0.0011 | +0.30% | <0.001 | +0.42% | <0.001 |
| 2023-2026 2023-01-01 → 2099-01-01 | 1,580,739 | +0.62% | <0.001 | +0.12% | <0.001 | +0.20% | <0.001 |
↓ Bearish triggers negative alpha = signal was right (stock underperformed market)
| Bench | Metric | 1d | 5d | 20d | 60d | 252d |
|---|---|---|---|---|---|---|
| spxew | Stock % | -0.04% | +0.20% | +1.18% | +3.23% | +12.49% |
| Bench % | +0.03% | +0.16% | +1.09% | +2.99% | +10.66% | |
| Alpha % | -0.07% | +0.03% | +0.08% | +0.20% | +1.80% | |
| Median alpha | -0.07% | -0.16% | -0.72% | -1.94% | -7.08% | |
| Hit rate (α>0) | 48.3% | 48.3% | 46.3% | 44.5% | 41.2% | |
| p (naive) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| p (HAC) | <0.001 | <0.001 | <0.001 | <0.001 | <0.001 | |
| N | 3,516,452 | 3,391,570 | 3,370,262 | 3,288,889 | 3,008,497 | |
| spx | Stock % | -0.04% | +0.20% | +1.18% | +3.23% | +12.49% |
| Bench % | +0.01% | +0.22% | +1.35% | +3.55% | +14.54% | |
| Alpha % | -0.05% | -0.01% | -0.13% | -0.30% | -2.06% | |
| Median alpha | -0.05% | -0.20% | -0.96% | -2.53% | -11.03% | |
| Hit rate (α>0) | 48.6% | 47.8% | 45.2% | 43.1% | 37.2% | |
| p (naive) | <0.001 | 0.0101 | <0.001 | <0.001 | <0.001 | |
| p (HAC) | <0.001 | 0.0563 | <0.001 | <0.001 | <0.001 | |
| N | 3,542,428 | 3,424,653 | 3,414,718 | 3,330,792 | 3,044,877 | |
| msci | Stock % | -0.04% | +0.20% | +1.18% | +3.23% | +12.49% |
| Bench % | +0.03% | +0.21% | +1.19% | +3.21% | +12.23% | |
| Alpha % | -0.06% | -0.00% | -0.01% | +0.09% | -0.00% | |
| Median alpha | -0.07% | -0.20% | -0.84% | -2.17% | -8.87% | |
| Hit rate (α>0) | 48.1% | 47.8% | 45.7% | 43.9% | 39.4% | |
| p (naive) | <0.001 | 0.6018 | 0.0353 | <0.001 | 0.9209 | |
| p (HAC) | <0.001 | 0.6989 | 0.2819 | 0.0052 | 0.9835 | |
| N | 3,529,517 | 3,410,352 | 3,384,231 | 3,311,081 | 3,026,706 |
Permutation null detail — all horizons × each benchmark
| Horizon | Bench | Observed lift | Null mean | Null 95% CI | pperm |
|---|---|---|---|---|---|
| 1d | spxew | +0.07% | +0.08% | [+0.08%, +0.09%] | 0.005 |
| 1d | spx | +0.08% | +0.09% | [+0.09%, +0.09%] | 0.005 |
| 1d | msci | +0.10% | +0.09% | [+0.09%, +0.10%] | 0.766 |
| 5d | spxew | +0.45% | +0.37% | [+0.36%, +0.38%] | 1.000 |
| 5d | spx | +0.45% | +0.39% | [+0.38%, +0.39%] | 1.000 |
| 5d | msci | +0.45% | +0.39% | [+0.38%, +0.40%] | 1.000 |
| 20d | spxew | +1.30% | +1.19% | [+1.18%, +1.21%] | 1.000 |
| 20d | spx | +1.33% | +1.22% | [+1.21%, +1.24%] | 1.000 |
| 20d | msci | +1.33% | +1.24% | [+1.22%, +1.25%] | 1.000 |
| 60d | spxew | +2.78% | +2.54% | [+2.52%, +2.56%] | 1.000 |
| 60d | spx | +3.02% | +2.61% | [+2.59%, +2.63%] | 1.000 |
| 60d | msci | +2.99% | +2.63% | [+2.61%, +2.65%] | 1.000 |
| 252d | spxew | +5.78% | +4.90% | [+4.86%, +4.95%] | 1.000 |
| 252d | spx | +6.02% | +5.25% | [+5.21%, +5.30%] | 1.000 |
| 252d | msci | +5.85% | +5.20% | [+5.17%, +5.25%] | 1.000 |
Example triggers on US large-caps (2023+, mcap ≥ $30B)
Six recent bearish NEW_HIGH_LOW triggers on US mega-caps, ranked within this sample of six: the top three and the bottom three, with extreme outliers excluded. Ranking is relative, not absolute — where the signal did well across the sampled names, even the bottom three can have beaten the benchmark, so the alpha printed on each panel is what settles it. Both groups are tail outcomes by construction; read them as the range, not the typical result.
Best three of the six sampled
Weakest three of the six sampled — not necessarily losses
Stock-regime quadrants (2×2 per-stock, 20d alpha detail table)
| Quadrant | N | Stock % (spxew) | Bench % (spxew) | Alpha % (spxew) | p (HAC) | Stock % (spx) | Bench % (spx) | Alpha % (spx) | p (HAC) | Stock % (msci) | Bench % (msci) | Alpha % (msci) | p (HAC) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Trending + Low vol Clean directional grind, low whipsaw | 339,418 | +0.01% | +0.88% | -0.80% | <0.001 | +0.01% | +1.13% | -1.07% | <0.001 | +0.01% | +0.98% | -0.91% | <0.001 |
| Trending + High vol Crisis selloff or parabolic rally | 1,171,420 | +2.13% | +1.22% | +0.88% | <0.001 | +2.13% | +1.53% | +0.65% | <0.001 | +2.13% | +1.33% | +0.79% | <0.001 |
| Non-trending + Low vol Quiet chop, summer doldrums | 485,147 | +0.04% | +0.88% | -0.78% | <0.001 | +0.04% | +1.14% | -1.05% | <0.001 | +0.04% | +1.01% | -0.92% | <0.001 |
| Non-trending + High vol Classical "whipsaw zone" for momentum | 1,647,129 | +1.09% | +1.15% | -0.04% | 0.0284 | +1.09% | +1.34% | -0.21% | <0.001 | +1.09% | +1.22% | -0.12% | <0.001 |
Sub-period breakdown table (20d alpha)
| Period | N | Alpha % (spxew) | p (HAC) | Alpha % (spx) | p (HAC) | Alpha % (msci) | p (HAC) |
|---|---|---|---|---|---|---|---|
| 2015-2019 2015-01-01 → 2020-01-01 | 1,094,023 | -0.15% | <0.001 | -0.24% | <0.001 | -0.02% | 0.2556 |
| 2020-2022 2020-01-01 → 2023-01-01 | 1,106,761 | -0.14% | <0.001 | +0.06% | 0.0114 | +0.20% | <0.001 |
| 2023-2026 2023-01-01 → 2099-01-01 | 1,441,394 | +0.43% | <0.001 | -0.18% | <0.001 | -0.17% | <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: one-sided fraction of 200 random-date null iterations falling in the “signal was right” tail (mean ≥ observed for bullish; mean ≤ observed for bearish). 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 not beaten random timing whatever the t-test says — and because the test is one-sided, a pperm up at its 1.000 ceiling is not "no edge" but inverted edge: every random draw served the claimed direction better than the trigger dates did.
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
Not a standalone entry trigger at 20 days. Bullish 20d alpha is +0.13% — inside the null : indistinguishable from random timing. Bearish 20d alpha is +0.08% — worse than random : firing on random dates would have done better. Fires are screening context inside a composite (section 4), not entries.
These verdicts are 20-day holds vs S&P 500 Equal Weight. Longer horizons can differ in either direction — check the permutation detail tables below before extrapolating.
2 · When it works — the setups that drive it
- Best bullish setup: Trending + High vol — alpha +0.30% / 20d on 1,397,458 historical triggers.
- Best bearish setup: Trending + Low vol — alpha -0.80% / 20d on 339,418 historical triggers.
- Best era for bullish: 2023-2026 — alpha +0.62% / 20d on 1,580,739 triggers.
- Best era for bearish: 2015-2019 — alpha -0.15% / 20d on 1,094,023 triggers.
3 · When it fails — common false positives
- Weakest bullish cell: Non-trending + Low vol — alpha +0.01% / 20d on 601,562 triggers.
- Weakest bearish cell: Trending + High vol — alpha +0.88% / 20d on 1,171,420 triggers.
- Worst era for bullish: 2015-2019 — alpha -0.37% / 20d on 1,357,451 triggers.
- Worst era for bearish: 2023-2026 — alpha +0.43% / 20d on 1,441,394 triggers.
Signal-specific failure patterns
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. Measured pair results — same-day co-fires put through the pair backtest — follow under “Measured pairings” below.
Breakout-family redundancy
New 20-day high, new 52-week high, and fresh 52-week high are breakout signals at different lookbacks — all fire when price exceeds the maximum of the prior N bars (Edwards & Magee, Technical Analysis of Stock Trends, 11th ed. 2018; Kirkpatrick & Dahlquist, Technical Analysis, 3rd ed. 2015; Bulkowski, Encyclopedia of Chart Patterns, 3rd ed. 2021). Stacking two or more in the same direction within a single Daily Report produces correlated rather than independent evidence.
Breakdown-family redundancy
New 20-day low, new 52-week low, and fresh 52-week low are breakdown signals at different lookbacks — all fire when price falls below the minimum of the prior N bars (Edwards & Magee, Technical Analysis of Stock Trends, 11th ed. 2018; Kirkpatrick & Dahlquist, Technical Analysis, 3rd ed. 2015; Bulkowski, Encyclopedia of Chart Patterns, 3rd ed. 2021). Stacking two or more in the same direction within a single Daily Report produces correlated rather than independent evidence.
Measured pairings — Bonferroni survivors
Beyond the literature pairings above, these are the same-day co-fire combinations involving 20-Day New High / New Low that cleared the pair backtest's Bonferroni cut on the full 2016–2026 sample (549 pairs × 5 horizons = 2,745 hypotheses), on universes filtered to ADV ≥ $5M, price ≥ $5 and market cap ≥ $100M. That cut is two-sided: it asks only whether the co-fire's α is reliably different from zero, in either direction, so a pair can survive by reliably underperforming — 12 of the 42 rows below do exactly that (negative full-sample α). The same run holds out 2023+: the Test columns are that held-out window, printed for every row with enough held-out co-fires to measure, so a survivor that did not repeat out of sample is visible rather than hidden. All α figures here are for holding the stock long after the co-fire — no shorting assumed, and no sign flip for bearish legs. So positive α means the co-fire was followed by outperformance and negative α by underperformance, whichever way either leg points — a bearish leg does not flip the reading. Survivors are rare by design — absence of a pair here means it did not clear the cut, not that it was untested. Ranked by held-out (2023+) α. Historical tendencies, not recommendations.
Pairs carry the underlying split names: new_20d_high is this signal's bullish side, new_20d_low its bearish side.
US (NYSE / NASDAQ / AMEX)
| Pair (same-day co-fire, long) | Full α | Test α (2023+) | Test N | p_perm test |
|---|---|---|---|---|
| new_20d_low bearish + weekly_change bearish | +0.61% | +1.06% | 25,567 | 0.002 |
| macd bearish + new_20d_low bearish | +0.55% | +0.85% | 9,573 | 0.002 |
| bullish_trend_breakdown bearish + new_20d_low bearish | +0.23% | +0.62% | 10,886 | 0.002 |
| bollinger bullish + new_20d_low bearish | +0.73% | +0.46% | 4,001 | 0.010 |
| hh_hl_streak bearish + new_20d_low bearish | +0.27% | +0.20% | 19,881 | 0.004 |
| double_top_breakout bullish + new_20d_high bullish | -0.56% | -0.22% | 4,680 | 0.164 |
| fresh_52w_high bullish + new_20d_low bearish | -3.37% | — | — | — |
7 of this universe's 18 surviving pairs involve this signal · α vs ^SPXEW.
Europe — 4 surviving pairs
| Pair (same-day co-fire, long) | Full α | Test α (2023+) | Test N | p_perm test |
|---|---|---|---|---|
| bullish_trend_breakdown bearish + new_20d_low bearish | +0.74% | +1.22% | 3,197 | 0.002 |
| new_20d_high bullish + volume_breakout bullish | +0.59% | +0.69% | 5,813 | 0.002 |
| hh_hl_streak bearish + new_20d_low bearish | +0.54% | +0.55% | 5,405 | 0.002 |
| macd bullish + new_20d_high bullish | +0.37% | +0.23% | 4,112 | 0.110 |
4 of this universe's 20 surviving pairs involve this signal · α vs ^STOXX.
Hong Kong — 6 surviving pairs
| Pair (same-day co-fire, long) | Full α | Test α (2023+) | Test N | p_perm test |
|---|---|---|---|---|
| macd bullish + new_20d_high bullish | +0.94% | +1.56% | 1,255 | 0.002 |
| new_20d_high bullish + volume_breakout bullish | +1.15% | +1.49% | 3,951 | 0.002 |
| hh_hl_streak bullish + new_20d_high bullish | +0.85% | +1.45% | 2,429 | 0.002 |
| bearish_trend_breakout bullish + new_20d_high bullish | +1.04% | +1.32% | 2,925 | 0.002 |
| new_20d_high bullish + weekly_change bullish | +0.89% | +1.05% | 5,551 | 0.002 |
| new_20d_low bearish + weekly_change bearish | +1.09% | +0.33% | 2,979 | 0.200 |
6 of this universe's 23 surviving pairs involve this signal · α vs ^HSI.
China A-shares — 25 surviving pairs
| Pair (same-day co-fire, long) | Full α | Test α (2023+) | Test N | p_perm test |
|---|---|---|---|---|
| new_20d_low bearish + weekly_change bearish | +3.22% | +3.68% | 50,447 | 0.002 |
| new_20d_low bearish + volume_breakout bearish | +1.32% | +3.60% | 9,979 | 0.002 |
| failed_double_top bearish + new_20d_low bearish | +2.36% | +2.20% | 1,860 | 0.002 |
| bollinger bearish + new_20d_low bearish | -2.26% | +2.00% | 147 | 0.032 |
| new_20d_low bearish + stochastics bullish | +1.37% | +1.91% | 16,120 | 0.002 |
| cci bearish + new_20d_low bearish | +0.96% | +1.47% | 881 | 0.002 |
| bearish_trend_breakout bullish + new_20d_high bullish | +0.70% | +1.43% | 38,486 | 0.002 |
| ma_crossover bearish + new_20d_low bearish | +1.19% | +1.38% | 1,372 | 0.002 |
| new_20d_low bearish + vwap_cross bearish | +0.91% | +1.17% | 3,909 | 0.002 |
| new_20d_low bearish + rsi bullish | +2.15% | +1.08% | 411 | 0.090 |
| hh_hl_streak bullish + new_20d_high bullish | +0.26% | +0.78% | 32,300 | 0.002 |
| new_20d_high bullish + stochastics bearish | +0.41% | +0.78% | 14,587 | 0.002 |
| macd bearish + new_20d_low bearish | +0.41% | +0.50% | 16,288 | 0.002 |
| double_bottom_breakdown bearish + new_20d_low bearish | +0.79% | +0.46% | 6,655 | 0.004 |
| hh_hl_streak bearish + new_20d_low bearish | +0.72% | +0.45% | 44,009 | 0.002 |
| new_20d_high bullish + volume_breakout bullish | -0.28% | +0.34% | 84,426 | 0.002 |
| new_20d_high bullish + weekly_change bullish | -0.56% | +0.15% | 83,483 | 0.010 |
| bollinger bullish + new_20d_low bearish | +0.58% | +0.02% | 7,647 | 0.836 |
| new_20d_high bullish + vwap_cross bullish | -0.75% | -0.21% | 12,388 | 0.078 |
| macd bullish + new_20d_high bullish | -0.61% | -0.55% | 18,940 | 0.002 |
| double_top_breakout bullish + new_20d_high bullish | -1.07% | -0.73% | 7,771 | 0.002 |
| new_20d_low bearish + williams_r bullish | -0.60% | -1.54% | 4,763 | 0.002 |
| new_20d_high bullish + volume_breakout bearish | -2.14% | -1.85% | 1,515 | 0.002 |
| new_20d_high bullish + vwap_cross bearish | -4.52% | — | — | — |
| new_20d_low bearish + rsi bearish | -3.37% | — | — | — |
25 of this universe's 138 surviving pairs involve this signal · α vs 83188.HK.
China A-share survivor α runs large but skews toward small-caps, where trading costs and thin liquidity claim a large share of any measured edge — screening context, not a capturable spread.
“—” in the test columns means the held-out 2023+ sample fell below the 20-observation minimum this run requires before it computes any statistic, so no out-of-sample figure exists for that pair — not that it never co-fired again. Those pairs rank last.
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 to cut the trigger noise — Fresh 20d highs are too common. Requiring volume above 1.5-2x the 20d average would reduce triggers by 60-80% and concentrate on real accumulation events.
- Combine with structural-breakout level — A 20d high that is also a 52w high or a multi-year resistance break is a structurally different event than a 20d high inside a sideways range. Compound filter should concentrate the alpha.
- Time-limit the hold to 20-30 days max — Past 20d, alpha reverts. A strict 20d time stop preserves the signal edge.
- Pair with trend filter — New 20d low in an uptrend is noise; in a downtrend it is continuation. 50DMA filter separates the two populations cleanly.
See also Why technical-only signals don't survive on their own for the broader argument.
5 · Before acting — a 5-point checklist
- 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.
- Where is price vs its own 50 / 200 DMA? A trend signal is only as credible as the underlying trend it claims to confirm. Check the 200DMA orientation before acting.
- 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.
- Is ADV20 enough for the intended size? A $1M notional order in a $500M name moves the tape by itself. A useful floor is adv20d ≥ 5% of the intended position.
- What invalidates the trade? Define a price level in advance (for longs: a close below the trigger-day low; for shorts: a close above the trigger-day high) and honour it. The backtest alpha is an average; any single trade can land at either tail.
Execution notes
Use primarily as a screen filter, not as a primary trigger. Entry open T+1 if used. Pair with volume + market-cap filters to cut the trigger rate substantially before evaluating any per-name decision. New lows side: Marginal short-side tile. Entry open T+1, exit within 20 trading days. Most useful as a universe filter for short candidates, paired with structural filters (50DMA, sector trend) for directional conviction.