Homing pigeon candlestick pattern: rules, examples and data

A homing pigeon is a two-candle pattern made of two down days. The first is a long black (red) candle. The second is a smaller black candle whose whole body sits inside the first one: it opens below the first day's open and closes above the first day's close. Sellers still won both sessions, but on the second day they could not even push price back to where the previous session ended.

That is the whole argument for the pattern: the decline is still running, yet its energy is shrinking. Traders who read it as bullish see the second, smaller red body as sellers running out of conviction near a low. It is close kin to the bullish harami, with one difference that matters: in the classic bullish harami the small second body is white, while in a homing pigeon it is still black.

This page covers how to recognize the pattern, the precise rules Finabase applies through TA-Lib's CDLHOMINGPIGEON function, why the second candle's color makes the signal weaker than many guides suggest, two real cases from Tesla and Intel, and what 416 homing pigeons on 60 large US stocks and ETFs did afterwards.

What a homing pigeon looks like

The pattern has two parts and one setting.

1. The first candle is long and black. It extends the decline with a session that opens high and closes much lower. A long body says sellers were in charge from the open to the close.
2. The second candle is small and black, nested inside the first body. It opens below the first candle's open, so there is no bounce at the open, and it closes above the first candle's close, so the selling that day stopped short.
3. The setting is a decline. The textbook pattern appears after several sessions of lower prices. Without a prior drop there is nothing for it to reverse.

The name is a picture: a bird returning to its nest, the second candle settling inside the first. Shadows are not part of the definition. Only the real bodies are compared, so the second day's high or low may poke outside the first body without changing anything.

The exact rules Finabase uses

Finabase's chart markers and scanner call TA-Lib's CDLHOMINGPIGEON. TA-Lib measures "long" and "short" against the stock's own recent candles rather than against fixed percentages. With the default candle settings the test is:

Three details deserve attention.

No trend filter. The TA-Lib source comment says the pattern is significant in a downtrend and then states plainly that the function does not consider the trend. A homing pigeon flagged in the middle of a rally is the same +100 as one at a low. In our sample 155 of 407 measured cases formed after a ten-session rise.

"Short" is generous. The second body only has to be no larger than the average body of the previous ten sessions, and that average includes the long first candle. After a big down day the bar for "short" rises, so a fairly ordinary red candle can qualify as the small one. In a volatile stock the second body can be a few dollars tall and still pass.

Equal opens or closes do not count. The comparisons are strict. If the second day opens exactly at the first day's open, TA-Lib returns 0. The textbook drawings rarely mention this, but it explains why a candle that looks right on a chart is sometimes not marked.

Why traders read it as bullish

Picture the two sessions as a negotiation. On day one sellers dominate the entire session. On day two they still win, since the close is below the open, but the territory they win is small and sits wholly inside yesterday's range of bodies. Buyers have not taken anything back. They have only stopped giving ground as fast.

The classic reading treats a shrinking black candle as a sign the decline is losing momentum, in the same family as the harami. A skeptic reads the same two days as a market that is simply still falling, at a slower pace. Both views fit the shape: a narrowing range after a drop can be the pause before a rebound or the pause before the next leg down. The candle cannot tell you which.

Confirmation before conviction

Because the second session still closed lower, a homing pigeon on its own carries no evidence of buying. Practitioners therefore look for proof in the sessions that follow:

• A close above the second candle's high. This is the rule in our test. It means buyers finally won a session outright.
• A white candle that closes above the first candle's midpoint. A stricter version that requires a real recovery into the long black body.
• Support underneath. A homing pigeon at a prior low, a rising long-term moving average or a round number has more to lean on than one in empty space.
• Momentum turning. An oscillator such as the Relative Strength Index (RSI): how to read and use it rising from below 30 while price makes the pattern adds a second, independent reading.

In the data below, confirmation made the largest difference of any filter we tried.

Two real homing pigeons

Tesla, June 6, 2025. On June 5, during a public dispute between Elon Musk and President Trump, TSLA opened at $322.49 and closed at $284.70, a 14.3% one-day drop and a body about five times the stock's average over the previous ten sessions. The next session opened at $298.83 and closed at $295.14: still a red candle, but entirely inside the previous body. The stock had fallen 13.5% over the ten sessions before. On June 9 it closed at $308.58, above the pigeon's high, which confirmed it, and ten sessions after the pattern TSLA was 18.1% higher. That is the version the textbooks draw: a violent down day, a hesitant follow-up and a rebound.

Intel, July 6, 2026. INTC had dropped 8.8% in ten sessions when it printed a long black candle on July 2 (open $128.92, close $120.35) and a 37-cent black body inside it on July 6. The shape was clean. The follow-through never came: the next session gapped down to close at $110.39, and ten sessions after the pattern the stock was 20.6% lower. The pigeon was never confirmed, since no close rose above its $127.30 high. A trader who required confirmation would have stood aside; one who bought the pattern blind took the full decline unless a stop was in place.

What 416 homing pigeons did next

We ran CDLHOMINGPIGEON over two years of daily bars (October 30, 2024 to October 29, 2026) for 60 large US stocks and ETFs, from AAPL and MSFT to TLT and XLF. The pattern appeared on every one of the 60 symbols, 416 times in all, roughly once every 72 trading days per symbol. For each case we measured the return from the pattern's close five and ten sessions later and compared it with every day in the sample.

(Of the 416 detections, 407 had ten sessions of later data and are measured.)

The raw pattern did not beat a random day. 50.9% of homing pigeons were higher five sessions later, a little worse than the 53.1% base rate in a market that mostly rose.

Context helped modestly. Pigeons that followed a ten-session decline, the setting the pattern is meant for, came in at 53.6%, while those formed after a rise were higher only 46.5% of the time. That gap is consistent with the textbook, though not large.

Confirmation was the real filter. About one pigeon in four was followed by a close above its high. Measured from that confirming close, so that the confirming day's own gain is not counted, 68.6% were higher five sessions later, with an average gain of 0.88%. That is one of the better confirmed results in the whole study, and a sample of 102 is large enough to take seriously, if not to trade on blindly.

Homing pigeon versus similar patterns

The homing pigeon and the Matching low candlestick pattern: equal closes as support often overlap. In our data 71 bars were flagged by both detectors, because a small black candle that closes just above the previous close can satisfy both rules at once.

A Hammer candlestick pattern: meaning, rules and reliability makes a different claim: within a single session, sellers pushed hard and buyers took most of it back. A homing pigeon never shows buyers winning anything, which is one reason it needs a confirming session even more than a hammer does.

An educational outline for a homing pigeon setup

This outline shows how a practitioner might structure a trade around the pattern. It is not a recommendation.

1. Check the setting. A decline of at least several sessions, ideally into a level that has mattered before.
2. Wait for a close above the second candle's high. In our sample this single condition separated the useful cases from the noise.
3. Put the invalidation below the pattern. The lower of the two lows is the natural line; a close beneath it means the decline resumed.
4. Size from that distance. Decide the dollar loss you accept first, then divide by the gap between entry and stop.

How Finabase shows homing pigeons

• Chart markers. Add the Homing Pigeon study to a chart (the link at the top of this page opens TSLA with it) and every candle where CDLHOMINGPIGEON returned +100 is marked.
• Scanner. The S&P 500 candlestick scanner sorts each detection by the sign of TA-Lib's output. The homing pigeon only ever returns +100, so it always appears on the Bullish list, including cases that formed in the middle of a rally.
• Signals tab gauge. On a symbol's Signals tab, each of the 61 candlestick patterns casts one vote: its most recent non-zero reading within the last 8 bars counts as Buy if positive and Sell if negative, otherwise Neutral. A homing pigeon in that window is always a Buy vote. The gauge then scores the group as (buy − sell) ÷ (buy + sell + the smaller of the neutral count and the larger side), reading Strong Sell below −0.6 and Strong Buy at 0.6 and above.

Frequently asked questions

Is a homing pigeon bullish or bearish?

It is classed as bullish, a potential reversal after a decline. Both of its candles are down days, though, so the bullish case rests on what happens next, not on the pattern.

How is a homing pigeon different from a bullish harami?

Only the color of the second candle. In the classic bullish harami the small inner body is white, meaning the second session closed up. A homing pigeon's is black, meaning sellers still won that day, which is why some traders rank it as the weaker of the two. TA-Lib's CDLHARAMI ignores the second candle's color, so all 416 homing pigeons in our sample were also flagged as bullish haramis.

Do the shadows have to fit inside the first candle?

No. TA-Lib and the classic definitions compare real bodies only. The second candle's shadows can extend beyond the first body without affecting the pattern.

Why does the Finabase scanner show a homing pigeon in an uptrend?

Because TA-Lib's function checks only the shape of the two candles and always returns +100. Judging whether a decline came first is left to the reader.

Summary

The homing pigeon pairs a long black candle with a smaller black candle tucked inside it, a picture of selling that continues but loses reach. On its own it told us little: across 416 cases on 60 large US names it was higher a week later about as often as any day. Once a later session closed above the pattern's high, the results improved markedly, to 68.6% of 102 cases. Use the shape to put a chart on your watchlist, and let confirmation, location and a predefined stop do the deciding.