Two crows candlestick pattern: how it forms and how often it fails

Two crows is a bearish three-candle pattern in which a rally gaps higher and then fails in two steps. A long white candle sets the trend. The next session opens above it but closes lower, a black "crow" perched above the white body. The third session opens inside that crow and falls all the way back into the white candle's body, closing the gap that the rally had just opened.

The crows are a traditional image of bad news circling overhead. In chart terms, the pattern says that a breakout above the white candle drew sellers instead of buyers, and that by the third day the market had given back most of the breakout.

This guide explains the formation, the rules TA-Lib applies, how it differs from its close relative Upside gap two crows: the bearish pattern that leaves its gap open, and the uncomfortable result of our test: in two years of data on large US stocks, two crows were followed by gains more often than by losses.

Anatomy of two crows

1. A long white candle extends an advance.
2. The first crow is a black candle whose body sits entirely above the white body. The open is higher than the prior close (the gap), but the session closes below its own open.
3. The second crow is another black candle. It opens within the first crow's body, not above it, and closes within the white candle's body.

Notice the two decisions the market makes. On day two it rejects the higher prices within a single session. On day three it does not even try to go higher at the open and sells off far enough to erase the gap.

The exact rules Finabase uses

Finabase uses TA-Lib's CDL2CROWS on daily bars. The function is short and strict:

Points worth noticing:

Only bodies count for the gap. The black candle's low may dip below the white candle's high. Many charting textbooks draw a full gap between candles; TA-Lib needs only a gap between bodies, which happens far more often in practice, sometimes by a few cents.

No size test on the crows. The first crow can be a stub or a long candle. Only the white candle has to be long.

No uptrend test. The function does not look at prices before the white candle. About two thirds (66.7%) of our detections followed a ten-session rise; the remainder formed in sideways or falling markets where the pattern has less meaning.

What the crows say about buyers and sellers

The white candle invites optimism. The next morning the stock opens higher, and for a few minutes the breakout looks like it is working. Then supply arrives. By the close the stock has fallen through its own opening price, leaving a black body above the gap. Buyers who chased the open are already losing.

The third morning, there is no fresh gap; the stock opens inside the prior black body and keeps falling until it is back inside the white candle's range. Everyone who bought above the white close in the last two sessions is now under water, and the gap that signaled strength has disappeared. That sequence of rejected breakout followed by a closed gap is the bearish story.

The story has a weak point, though. A strong uptrend often digests a gap this way and then resumes, especially when the white candle itself was a big earnings-day move. The pattern describes a two-day pullback; whether it becomes more depends on what comes next.

Confirming two crows

• A close below the white candle's open, which fully reverses the long white day.
• A close below the second crow's low, the rule used in our statistics.
• Volume expanding on the crows rather than on the white candle.
• Overhead resistance at the level where the first crow failed.

Seven of the 31 detections in our sample confirmed by the second rule, which shows how often the market simply bounced instead.

Two real cases

Disney, May 8, 2026. DIS jumped on May 6, opening 6% above the prior close at $106.52 and finishing at $108.06 for a white body of $1.54. On May 7 it opened higher again at $109.05 but closed at $108.66, a black body just above the white one. On May 8 it opened inside that crow at $109.03 and closed at $108.02, a few cents inside the white body. The next session closed at $104.72, below the pattern's low, and ten sessions later the stock was 4.6% below the May 8 close. Here the crows marked exactly where the gap-up enthusiasm ran out.

AMD, April 20, 2026. AMD had risen 24.9% in ten sessions when a $13.27 white body on April 16 was followed by a black candle on April 17 whose body cleared the white close by only 13 cents, and a second black candle on April 20 that closed at $274.95, inside the white body. Every rule was met. The next session closed at $284.49, back above the second crow's open, and ten sessions later AMD was 24.2% higher. The tiny body gap is a reminder that TA-Lib's version of the pattern can be far less dramatic than the drawings in books.

What the data says

Across 60 large US stocks and ETFs, two years of daily bars (October 30, 2024 to October 29, 2026), CDL2CROWS fired 31 times on 28 symbols. One case was too recent for a full five-session follow-up.

The direction is the opposite of the textbook. Two thirds of the crows were followed by higher prices a week later, and the average five-session gain was nearly ten times the baseline. A few very large rebounds, such as AMD's, inflate that average: three of the detections were followed by five-session gains of 19% to 31%.

Two crows versus similar patterns

The cleanest way to keep the two crows family straight: in two crows the third candle opens inside the first crow and closes inside the white body; in upside gap two crows the third candle opens above the first crow and stops above the white body. TA-Lib codes them as separate functions, and one formation never satisfies both.

An educational outline

This shows how the pattern is approached in practice. It isn't a recommendation to trade.

1. Check the white candle. A huge white body on news, like Disney's, sets up a gap that the market may want to retest. A modest one in a steady trend is weaker context.
2. Wait for a close below the pattern's low. Our data suggests the unconfirmed pattern is not a reliable short signal on large caps.
3. Place invalidation above the first crow's high. A close there means the breakout resumed.
4. Size from that distance. The space between the confirmation close and the first crow's high is the risk per share.

How Finabase shows two crows

• Chart markers. The Two Crows study marks the second crow, the third candle of the pattern, wherever TA-Lib finds it on daily bars.
• Scanner. The pattern is always scored −100, so stocks completing it today appear on the bearish list only.
• Signals tab. In the Candlestick gauge on a page such as AAPL's signals, two crows within the last eight bars casts a Sell vote among the 61 patterns. The gauge score is (buy − sell) / (buy + sell + min(neutral, max(buy, sell))); below −0.6 reads Strong Sell, below −0.2 Sell, below 0.2 Neutral, below 0.6 Buy, and above that Strong Buy.

Frequently asked questions

What is the difference between two crows and three black crows?

Two crows is a failed gap above one long white candle; it ends inside that candle's body. Three black crows is three long black candles in a row, each closing lower, and it does not need a gap (TA-Lib's version does check for one white candle just before it).

Does the gap have to be a full gap with no overlapping shadows?

Not in TA-Lib. Only the real bodies must not overlap. AMD's April 2026 example cleared the white close by 13 cents.

Why did two crows do so badly in the test?

Most detections formed during strong advances in large, heavily traded stocks, where two-day pullbacks were routinely bought. With only 30 measurable cases, a few sharp rebounds also dominate the averages.

Is two crows more bearish than upside gap two crows?

By the textbook logic, yes, because the gap is closed. In our small samples the upside gap version did better as a bearish signal (55.6% lower after five sessions, against 33.3%), but with 18 and 30 cases neither sample is large enough to rank them.

Summary

Two crows records a breakout above a long white candle that fails twice: once within the session after the gap, and again the next day when prices fall back into the white body. The logic is sound, and Disney's May 2026 case played out as described. Yet on 60 large US names over two years, the 30 measurable detections were followed by gains two times out of three. Treat the pattern as the start of a pullback that still needs proof, wait for a close below its low, and know exactly where a resumed rally would prove it wrong.