Three white soldiers (three green candles in a row) mean it keeps rising
Many traders read three strong green candles in a row as momentum that should continue. This rule records the pattern and compares the 20-trading-day outcome with similar stocks on the same day, without recommending a trade.
Result
No difference — The test did not establish a difference from the matched baseline.
Why this verdict · no difference in rise rate vs similar-volatility stocks was confirmed
- Observed up-rate
- 48.8%
- Volatility-matched expected rate
- 50.6%
- 95% lift interval
- 0.88 ~ 1.04x
- Signals · symbols · dates
- 890 · 824 · 207
As of 2026-10-23 · Observed period 2025-10-21 to 2026-10-23
Source: DawnScan daily scan records and daily price records
Supplementary metrics
By period
Supplementary metrics · the preregistered verdict uses 20 trading days · viewing several periods can make some cells look favorable by chance
| Period | Observed | Expected | Difference (pp) | Difference 95% interval | Sample |
|---|---|---|---|---|---|
| 5 trading days | 49.0% | 50.1% | -1.1 | -4.4 to 2.5 pp | 957 / 965 · no future price 8 |
| 10 trading days | 50.2% | 50.7% | -0.6 | -4.0 to 2.8 pp | 933 / 965 · no future price 32 |
| 20 trading days | 48.8% | 50.6% | -1.9 | -5.9 to 1.8 pp | 890 / 965 · no future price 75 |
| 40 trading days | 50.4% | 50.9% | -0.5 | -3.6 to 2.5 pp | 843 / 965 · no future price 122 |
| 60 trading days | 51.1% | 51.3% | -0.2 | -3.7 to 3.4 pp | 803 / 965 · no future price 162 |
By entry-price range · 20 trading days
Supplementary metrics · the preregistered verdict uses 20 trading days · viewing several periods can make some cells look favorable by chance
| Entry price | Observed | Expected | Difference (pp) | Difference 95% interval | Sample |
|---|---|---|---|---|---|
| Under $5 | 45.3% | 45.1% | 0.2 | -7.0 to 7.0 pp | 137 / 153 · no future price 16 |
| $5–$20 | 41.9% | 48.9% | -7.0 | -13.1 to -1.3 pp | 308 / 336 · no future price 28 |
| $20 or more | 54.6% | 53.5% | 1.1 | -4.6 to 6.7 pp | 445 / 476 · no future price 31 |
Rule definition
An event is the day a three-white-soldiers pattern completes: three green candles in a row, three rising closes, each body at least 60% of its high-low range, and each open inside the previous body. No prior decline is required.
Why people use this rule
Many traders read three strong green candles in a row as momentum that should continue. This rule records the pattern and compares the 20-trading-day outcome with similar stocks on the same day, without recommending a trade.
Results by market regime
| Regime | Cases | Observed | Expected | Verdict |
|---|---|---|---|---|
| Neutral | 203 | 34.5% | 37.4% | No difference |
| Risk-off | 75 | 74.7% | 70.9% | Not enough data |
| Risk-on | 611 | 50.2% | 52.5% | No difference |
| Other | 1 | 100.0% | 79.3% | Not enough data |
Verdict history
Solid line: observed rate · dashed line: volatility-matched expected rate.
Recent confirmed cases
| Date | Symbol | Outcome |
|---|---|---|
| 2026-10-23 | UIS | -1.6% |
| 2026-10-23 | OZK | -5.3% |
| 2026-10-23 | MGF | -0.8% |
| 2026-10-23 | GDYN | +46.4% |
| 2026-10-22 | IVA | -4.2% |
| 2026-10-22 | INFY | -11.9% |
| 2026-10-22 | EPAM | +8.5% |
| 2026-10-22 | DBL | -4.9% |
| 2026-10-22 | CBNA | +1.8% |
| 2026-10-22 | BVS | -33.8% |
| 2026-10-21 | PATH | -12.2% |
| 2026-10-21 | IBM | -8.4% |
| 2026-10-20 | DGNX | -23.2% |
| 2026-10-20 | CQP | -6.2% |
| 2026-10-15 | SWKS | +13.8% |
| 2026-10-15 | FRHC | -12.4% |
| 2026-10-15 | ENGS | -86.6% |
| 2026-10-14 | TG | -25.0% |
| 2026-10-13 | YMM | +1.9% |
| 2026-10-12 | ONON | +6.2% |
| 2026-10-12 | GME | -6.0% |
| 2026-10-12 | DKS | +7.5% |
| 2026-10-12 | CNI | +7.3% |
| 2026-10-12 | AXIN | +0.5% |
| 2026-10-09 | THC | +7.4% |
| 2026-10-09 | TDUP | -48.2% |
| 2026-10-09 | NAVN | +7.8% |
| 2026-10-09 | INGN | -2.4% |
| 2026-10-09 | AXP | +4.2% |
| 2026-10-08 | SKWD | +7.7% |
| Date | Symbol | Outcome |
|---|---|---|
| 2026-10-23 | UIS | -1.6% |
| 2026-10-23 | OZK | -5.3% |
| 2026-10-23 | MGF | -0.8% |
| 2026-10-23 | GDYN | +46.4% |
| 2026-10-22 | IVA | -4.2% |
| 2026-10-22 | INFY | -11.9% |
| 2026-10-22 | EPAM | +8.5% |
| 2026-10-22 | DBL | -4.9% |
| 2026-10-22 | CBNA | +1.8% |
| 2026-10-22 | BVS | -33.8% |
| 2026-10-21 | PATH | -12.2% |
| 2026-10-21 | IBM | -8.4% |
| 2026-10-20 | DGNX | -23.2% |
| 2026-10-20 | CQP | -6.2% |
| 2026-10-15 | SWKS | +13.8% |
| 2026-10-15 | FRHC | -12.4% |
| 2026-10-15 | ENGS | -86.6% |
| 2026-10-14 | TG | -25.0% |
| 2026-10-13 | YMM | +1.9% |
| 2026-10-12 | ONON | +6.2% |
Show more (10)
| Date | Symbol | Outcome |
|---|---|---|
| 2026-10-12 | GME | -6.0% |
| 2026-10-12 | DKS | +7.5% |
| 2026-10-12 | CNI | +7.3% |
| 2026-10-12 | AXIN | +0.5% |
| 2026-10-09 | THC | +7.4% |
| 2026-10-09 | TDUP | -48.2% |
| 2026-10-09 | NAVN | +7.8% |
| 2026-10-09 | INGN | -2.4% |
| 2026-10-09 | AXP | +4.2% |
| 2026-10-08 | SKWD | +7.7% |
Method and limitations
Definition: An event is the day a three-white-soldiers pattern completes: three green candles in a row, three rising closes, each body at least 60% of its high-low range, and each open inside the previous body. No prior decline is required.
Historical outcomes only. The sample may reflect a limited market regime. Results do not account for stops, position sizing, fees, or taxes.
Results by data period
| Period | Dates | Cases | Observed rate (primary metric) | Expected (same day, same volatility) | Ratio (95% interval) | Delisted cases | Sample |
|---|---|---|---|---|---|---|---|
| Reconstructed | 2025-10-21 – 2026-06-12 | 646 | 51.1% | 52.7% | 0.97x (0.88–1.05) | 0 | Sufficient |
| Daily records | 2026-06-16 – 2026-10-23 | 244 | 42.6% | 45.1% | 0.95x (0.77–1.13) | 2 | Sufficient |
The reconstructed period uses past prices of stocks listed at the time, so stocks delisted before then may be missing (delisted cases in this period: 0).
Down-market (risk-off) cases number 75 in the reconstructed period and 0 in the daily-record period, and the two periods also differ in market-regime mix, so the gap cannot be attributed entirely to the data range. The verdict is one judgement on the combined data.
Data range: results use daily records we collected ourselves from 2026-06-16; earlier periods are reconstructed from past prices of stocks listed at the time (stocks delisted before then may be missing).
Past data · not investment advice