Concealing Baby Swallow Candlestick Pattern: Rules, Trading and Backtest

The concealing baby swallow is a four-candle bullish reversal pattern made only of black candles. Two black marubozus extend a downtrend, a third black candle gaps down but trades back up into the second body, and a fourth black candle swallows the third whole, shadows included. The third candle is the "baby" that gets concealed.

It is extremely rare. This guide gives the exact rules, the reading behind them, how to trade and backtest it, and what our test bench found when it scanned full history on stocks, crypto, forex, ETFs, indices and commodities: not a single occurrence under its strict rules.

What is the concealing baby swallow?

It is a bottom reversal signal described by Gregory Morris in Candlestick Charting Explained and implemented in the open-source TA-Lib library as the CDLCONCEALBABYSWALL function. What makes it unusual is that every candle is black: the bullish reading comes from how the selling behaves, not from a strong up day.

Setups such as the bullish engulfing or the morning star end with a candle in the new direction. This one asks you to see exhaustion inside a decline that still looks intact. That is also why it is hard to trade.

Identification and structure of the pattern

The four-candle formation

1. First candle: a black marubozu after a downtrend, a long body with no upper or lower shadow (or almost none).
2. Second candle: a second black marubozu. Our bench rule, like the TA-Lib function, does not require it to close below the first.
3. Third candle (the baby): black, it opens with a gap below the previous close, but its upper shadow reaches back up into the second body.
4. Fourth candle: tall and black, it opens above the top of the baby and closes below its bottom, so it covers the whole baby, shadows included.

Visual characteristics and rules

The chart below is a schematic drawing of the rule, with the faded bars showing the decline before it and the lines of the position we test: entry above the high, stop below the low, target at twice the risk. Our bench found no real occurrence under the current rule; the section Real example: why our data has none explains why.

Our bench applies the rule without judgment: after a downtrend (the close before the setup is at least one ATR below the close ten bars earlier and in the lower half of the last 40 closes), two black marubozus; a black baby opening with a gap down whose upper shadow reaches into the second body; a black fourth one that covers the baby, shadows included. "Long" and "very short shadow" are measured against the 10 prior candles, as in TA-Lib, but a long body must exceed 1.3 times the average body, not 1.0, and 0.6 times the largest of those 10 bodies.

Differences from similar reversal patterns

• The ladder bottom also starts with black bars stepping lower, but it ends with a strong up day. The concealing version never does.
• The homing pigeon is two black bars, the second inside the first body. Here the fourth one is the large one and the baby sits inside it.
• The matching low is two black bars closing at the same level, another sign of selling that stops making progress.
• A bullish engulfing needs an up day covering a down day. Here what covers the baby is itself black, which sets this formation apart from the usual reversal setups.

Market psychology behind the formation

The first two marubozus show sellers in full control: price opens at the top and closes at the bottom, two candles in a row. The baby opens with a gap down, which looks like more of the same, but buyers push it back up into the previous body before sellers close it lower again. That shadow is the first sign that cheaper levels attract demand and that the power of the bears is fading.

On the fourth day, price opens above the top of the baby: the gap down has already been erased at the open. Sellers still force a close under the baby, but they need much more power and a much larger range to do it. The classic reading is a selling climax: the last holders give up, and there is little supply left below. The story is coherent, but it rests on a four-day sequence we can almost never observe, so there is no data to check it.

How to identify the pattern on your charts

Step-by-step detection criteria

1. Confirm the prior trend is down: the close before the setup is clearly below the close 10 bars earlier and in the lower half of its recent range.
2. Check that bars 1 and 2 are black with almost no shadows.
3. Check the gap: the baby must open below the second close.
4. Check the upper shadow of the baby: its high must be above the second close, so that shadow and body overlap.
5. Check the fourth: open above the high of the baby, close below its low.
6. Wait for the close of the fourth before marking anything.

Common mistakes to avoid when spotting the pattern

• Accepting candles with visible shadows as marubozus, which need long bodies with almost no wick. The rule is strict, and loosening it gives a different, untested setup.
• Ignoring the gap. Without the gap down when the baby opens, it is not the same formation.
• Calling a fourth candle that only covers the body of the baby a valid signal. It must cover the shadows too.
• Expecting an up day inside the setup. None of the four bars is green or white; a white fourth candle makes it a different signal.

Trading strategies using this candlestick signal

The trade below is the one our bench simulates on every reversal setup. It is a framework for your own tests, not a proven strategy for this formation.

Setting up your entry point

Place a buy stop order just above the high of the four bars and wait for price to trade through it. Because the setup ends with a black candle at a new low, buying at the close of the fourth day means buying a stock that is still falling: a stop order above the high asks the market to prove the reversal first.

Determining effective stop-loss placement

Profit targets and risk-to-reward ratios

Our bench sets the profit target at twice the risk (a 2:1 reward-to-risk ratio). The 20-day exponential moving average (EMA) above price, a declining trend line, or the last swing high are alternative profit targets. You can also take partial profit at the first target and trail the stop loss for the rest. Decide the profit target before entry and record every trade, profit or loss.

Reliability and performance of the pattern

We cannot tell you how reliable it is, because there is not enough data to measure it. Our test bench scanned the full history of every asset we cover with the rule above and found no occurrence at all: none on stocks, crypto, forex, ETFs, indices or commodities, and none on hourly, 4-hour, daily or weekly charts. We report it as absent, not as a zero win rate. The strict definition explains part of it: checked against TA-Lib on candle geometry alone, before our trend filter and our longer bodies, our rule keeps 1 of the 8 cases TA-Lib flags (12.5%).

Why it is considered a rare bullish reversal signal

Each condition is already uncommon, and the setup needs all of them in a row. For scale, our bench counted 135,183 black marubozus on daily charts, but not one sequence that met the full rule. Two marubozus back to back are rare, a gap down with an upper shadow back into the prior body is rarer, and a fourth black one that opens above that shadow and closes below the low is rarer still.

Historical win rates and statistical significance

We publish a win rate only when a setup has at least 30 occurrences, and we compare it with random dates on the same assets. With no case at all, there is no percentage to compute. Before trusting a success rate quoted for this formation, check how many cases it rests on.

The closest tested relatives give a sense of what to expect from a bullish setup built on black bars. The ladder bottom (44 daily occurrences) was followed by an increase 10 days later 54.5% of the time, against 54.3% for random days on the same assets, and its 2:1 trade, opened on only 23 of them, reached the profit target 21.7% of the time against 35.1% on random dates, too few trades to judge. The homing pigeon (2,432 daily occurrences, 901 closed trades) reached its 2:1 target 34.3% of the time against 34.9% on random dates (+0.03 R against +0.07 R, not a significant gap). On the trade, the homing pigeon shows no clear edge, and the ladder bottom has too few trades to judge.

Real example: why our data has none

We cannot show a real chart, because no bar in our data passes the current rule. An earlier, looser version of our bench flagged a single daily bar on a US stock in September 1987. Since then we tightened two conditions: a marubozu needs a body clearly longer than the recent average, and the prior decline must be a real one, not a short pullback inside a rising market. That 1987 sequence no longer passes.

Each tightening comes from a blind reading of real charts, where readers rejected setups with small bodies or no prior trend. Applied to a four-candle formation that already needed two marubozus, a gap and a full cover, they leave nothing. If you scan with TA-Lib, expect a few hits it accepts and we do not: they are looser versions, not the formation tested here. One case would have proved nothing anyway.

Technical filters

The usual filters are listed below. We have not tested any of them on this formation, since there is nothing to test them on.

Using volume to confirm the reversal

The selling climax reading predicts heavy volume on the baby and the last day. A volume filter would ask for an increase in volume on the last day, above its recent average volume, as a sign that sell orders were absorbed. For the ladder bottom, volume above 1.5 times the average showed no benefit in our data: 11.1% of 9 trades reached the 2:1 target, against 28.6% of 14 on normal volume. These samples are too small to conclude and the split is descriptive, not corrected for multiple testing, which is a reason to test a volume rule before trusting it.

Combining with momentum indicators

Momentum indicators add one more condition. Possible choices: the default 14-period RSI below 30, a bullish divergence on the RSI or MACD indicator (price makes a new low while momentum does not), or a fast EMA (exponential moving average) that crosses back above a slow EMA after the signal, a moving average cross that confirms the end of the decline. Each filter adds a reason for the trade, but it also cuts the number of signals, which is already close to zero here.

The role of support and resistance levels

A setup that forms on a known support zone, a prior swing low or a long-term moving average such as the 200-day line, is read in classical analysis as carrying more context than one in the middle of a trend; we have not tested that on this formation. On the ladder bottom, 59.4% of daily setups below the 200-day moving average were higher 10 sessions later; only 12 setups sat above it, too few to read. That breakdown is descriptive, not corrected for multiple testing.

Pros and cons of trading the concealing baby swallow

Advantages for swing and day traders

• Precise, objective rules that a scanner can check.
• A natural stop loss below the low of the setup.
• It describes selling exhaustion, a useful idea even when the full formation is absent.

Potential risks and limitations

• It is extremely rare: no occurrence in our whole database under the strict rule.
• No statistic can confirm or deny its reliability.
• It ends at a new low, so entering early means buying a falling market.
• On intraday charts of markets that trade around the clock, the gap condition almost never happens.

Backtesting the concealing baby swallow

Importance of systematic testing

A setup this rare cannot be judged from a few charts picked by eye. A proper backtest applies the same rule to every asset and every bar, counts every occurrence, wins and losses alike, and compares the result with the same trade on random dates. That is the method our bench uses, described in our methodology.

Key parameters for algorithmic backtesting

• Input data: open, high, low, close for each bar, adjusted for splits, on as many stocks as you can.
• Trend filter: how you define the prior downtrend (we use the close 10 bars earlier, one ATR lower at least, and the range of the last 40 closes).
• Thresholds: what counts as a "very short" shadow and a "long" body, relative to recent bars.
• Trade rules: entry above the high, stop loss below the low, profit target as a multiple of risk, and a time limit.
• Timeframe: start with daily bars, where gaps exist.
• Sample size: at least 30 occurrences before reading a win rate, and a correction when you backtest many setups at once.

In Python, the TA-Lib call takes the open, high, low, close arrays: talib.CDLCONCEALBABYSWALL(open, high, low, close). It returns 100 on the bar that completes the setup and 0 elsewhere. Run it on a single stock and expect a column of zeros.

Frequently asked questions

Is the concealing baby swallow bullish or bearish?

Bullish. It is a bottom reversal signal that appears after a downtrend, even though all four bars are black.

Is it safe to trade this pattern in isolation?

No. With no occurrence in our data, there is no evidence that it works on its own. Treat it as context and always use a stop.

What is the best timeframe for this pattern?

Daily, because it needs a gap at the open, and gaps mostly happen between trading days. The one case an earlier, looser rule found was on a daily stock chart.

Can this pattern be used for forex and stock trading?

In theory, yes. In practice our strict rule finds it nowhere, and it is least likely on forex, where the market trades around the clock and rarely gaps.

Conclusion: is it right for your strategy?

Probably not as a stand-alone signal, and nothing on this page is financial or investment advice. It is a precise definition of selling exhaustion, but our bench found no occurrence in the full history of every market it covers once the rule was made strict. If you see one, treat it as a location for a trade, not as a reason for one: wait for price to break above the high, keep the stop loss below the low, and keep the risk small.

About the author and the data. Written by Amaury, founder of PatternsWizard, who is passionate about finance, quantitative trading and data analysis. He built the PatternsWizard test bench, which detects each setup with published rules on full market history: crypto, forex, US stocks, ETFs, indices and commodities. Trades and win rates are compared with random bars on the same assets, excess returns with the rest of the market, with error margins and a correction for multiple testing, and the whole pipeline is checked on random-walk prices, where nothing can be predicted. For this formation the bench found no occurrence, so it publishes no statistics; the counts above come from the run of November 5, 2026. The full methodology explains every step.