Guide · 3 plays
Candles & chart patterns: what it predicts, and when it works
A short pattern of bars can show a move running out of steam. Pattern plays look for a turn in the last day or few days, like a candle that swallows the one before it or a breakout that fails and snaps back, and trade the reversal.
Why it might work
A failed move traps the traders who bet on it. When a break below an old low reverses the next day, or a big up candle erases a down day, those traders rush to get out and add fuel to the turn.
When it fails
Patterns are common and most of them mean nothing: the same candle shape appears in random data. In a strong trend, a reversal pattern is usually just a pause before the trend carries on.
What the research says
- Steve Nison's Japanese Candlestick Charting Techniques (1991) brought candle patterns, used in Japanese rice trading since the 1700s, to Western traders.
- Caginalp and Laurent (1998) found some three-day candlestick reversal patterns predicted S&P 500 stocks over 1992 to 1996.
- Marshall, Young and Rose (2006) tested candlestick rules on the Dow stocks from 1992 to 2002 and found they added no value after accounting for chance.
- Lo, Mamaysky and Wang (2000) measured classic chart patterns with statistics and found they carried some information about the next few days' returns, mostly small.
Every candles & chart patterns play, ranked
Share of stocks where each play did better than holding, since 2005 or as long as the stock has traded, after costs. Click a column to sort.Sort with the menu above the list.
| Play | Beat holding, risk-adjusted | Cut the worst fall | Beat holding on return | Calls right | Switches / yr | Time in | In now |
|---|---|---|---|---|---|---|---|
| HikkakeDaniel L. Chesler | 18% | 93% | 15% | 35% | 4.4 | 33% | 101/377 |
| Engulfing SwitchSteve Nison & Thomas N. Bulkowski | 12% | 87% | 11% | 50% | 6.2 | 44% | 169/377 |
| Turtle Soup Plus OneLinda Bradford Raschke & Larry Connors | 7% | 100% | 10% | 44% | 2.5 | 2% | 13/377 |
When it worked
The family did best against holding in the 2008 crash (+26% median gap) and worst in the 2020–21 rally (−34%). Median gap in return against holding (per year over stretches longer than 18 months; crypto excluded).
| 2008 crash | Recovery 2009–19 | COVID drop | 2020–21 rally | 2022 bear | Since Oct 2022 | |
|---|---|---|---|---|---|---|
| Candles & chart patterns | +26% | −12% | +21% | −34% | +11% | −10% |
Where it worked
By kind of stock: how often the family's plays beat holding on risk-adjusted return, and the median Sharpe ratio gap.
| Group | Beat holding, risk-adjusted | Sharpe gap |
|---|---|---|
| Real estate | 19% | -0.26 |
| Micro caps | 17% | -0.41 |
| Industrials & rentals | 15% | -0.27 |
| Software & devices | 15% | -0.29 |
| Energy & power | 15% | -0.26 |
| Industrials & materials | 14% | -0.30 |
| Metals & mining | 13% | -0.26 |
| Sectors | 11% | -0.31 |
| Financials | 10% | -0.27 |
| Consumer & retail | 10% | -0.33 |
| Indexes & ETFs | 10% | -0.27 |
| Healthcare | 9% | -0.28 |
| Media & telecom | 3% | -0.30 |
| Semiconductors | 3% | -0.41 |
| Big tech | 0% | -0.57 |
How traders use it
- Treat a pattern as a reason to look, not a reason to trade: the evidence that patterns alone pay is thin.
- Pattern plays hold for days, not months, so trading costs and slippage matter a lot.
- A pattern that agrees with the longer trend (a bullish turn inside an uptrend) has a better record than one that fights it.
Right now
The candles & chart patterns plays hold 25% of the stock-and-play combinations Be The Puck tracks (283 of 1,131) as of the Oct 13, 2026 close.
See which stocks each family holds · The plays and their rules
This analysis does not constitute trading advice. Please meet with an advisor or independently review sources before making any decision.