The pattern family tree: harami confirmed = three inside up; engulfing confirmed = three outside up. How a third candle converts a tentative two-candle signal into a confirmed reversal — and the bearish mirrors.
These extend the two-candle vocabulary by adding a confirmation candle — converting tentative signals into structurally proven reversals.
Three inside up, three inside down, three outside up, three outside down, and concealing baby swallow — structure showing the two-candle parent pattern plus its confirmation candle
| Pattern | Confirmation | Volume | Reliability | Common Failure Mode |
|---|---|---|---|---|
| Three inside up / three inside down | Third candle is the confirmation built into the pattern. Further follow-through strengthens. | Ideal — light volume on the harami middle candle, expanding volume on the third confirming candle. | Generally regarded as more reliable than the bare harami because confirmation is structurally required by the pattern definition. | Third candle is short and doesn't decisively break the prior trend. |
| Three outside up / three outside down | Third candle is confirmation. Follow-through next session strengthens. | Expanding volume on the engulfing candle (second) and the confirmation candle (third) is the strongest signal. | Generally regarded as more reliable than a bare engulfing pattern because confirmation is structurally required. | Same as engulfing — occurs in sideways chop with no real trend to reverse. |
| Stick sandwich | Next candle closes above the pattern's high (for bullish stick sandwich). | No specific signature widely emphasized; volume confirmation on the breakout candle is the key check. | Generally regarded as less reliable than morning star or three inside up. More of a curiosity pattern in most curricula. | Frequent. The two matching closes can be coincidence rather than meaningful support. |
| Matching low | Next candle closes above the matching low level. | Heavier volume on the second matching candle suggests real seller exhaustion at the level. | Modest. Better treated as a level-identification tool than a trade signal. | The level breaks on the third or fourth session. |
| Concealing baby swallow | Next candle continuing higher, ideally with a gap up. | Heavy volume on the fourth (engulfing) candle is the key signature. | Rare enough that meaningful reliability data is hard to come by. Generally regarded as a high-conviction signal when correctly identified, but verification is difficult. | Misidentification is the dominant failure — the pattern's specific structural requirements (two marubozu, gap-down third, engulfing fourth) are easy to approximate but hard to find exactly. |
Three inside up / three inside down and three outside up / three outside down are covered in the standard multi-source pools but specific aggregate statistics are less widely cited than for single and two-candle patterns, partly because their structural requirements are strict enough that sample sizes are smaller in backtests. The four primary sources students can consult: thepatternsite.com (Bulkowski — post-breakout 10-day performance), liberatedstocktrader.com (Moore — 10-day holding period on Dow components), quantifiedstrategies.com (independent SPY-based backtest), and academic literature via Google Scholar and SSRN. Teaching point: for three inside up and three outside up, the reliability advantage over the bare parent patterns (harami, engulfing) is the central claim — not a specific percentage. The structural confirmation requirement filters out the ambiguous harami and engulfing instances that fail quickly, leaving a population of signals that by definition have already produced a confirming third candle. Students who understand this filtering mechanism understand why the three-candle versions outperform the two-candle versions without needing a specific number. Stick sandwich and matching low: generally treated as level-identification tools in the literature rather than high-reliability trade signals. Both patterns identify a closing price that held twice; the question is whether the market will defend it a third time. Concealing baby swallow: rarity itself reduces the available backtest sample. Pattern sites note the pattern's extremely strict structural requirements — two marubozu, gap-down third, engulfing fourth — make random false positives extremely unlikely, which is the qualitative basis for the high-conviction assessment. Multi-source pool: thepatternsite.com, liberatedstocktrader.com, quantifiedstrategies.com, litefinance.org, patternswizard.com.
Key Takeaways
A student sees a bullish harami in a downtrend. The next candle is a small doji — not clearly bullish. Is this a completed three inside up?