Unique three river bottom: rules, real examples and test results

The unique three river bottom is a slow-motion bottoming pattern. A long black candle extends a decline. The next session's black body is smaller and sits inside the first one, yet it probes to a new low before recovering. The third session is a small white candle that stays above that new low. Selling reached a fresh extreme on day two and was rejected; day three is the first quiet session in which buyers close the stock higher than it opened.

The name is a literal translation from the Japanese tradition, and the pattern is one of the rarer ones Nison describes. Its logic resembles a Hammer candlestick pattern: meaning, rules and reliability stretched over three sessions: the long lower shadow of the second candle is the rejection, and the white third candle is the beginning of the response.

This page explains the formation, shows how TA-Lib's version is considerably looser than the textbook one, walks through a case in Intel and one in Home Depot, and reports the record of 44 detections in two years of large-cap data.

Anatomy of the unique three river bottom

Three candles, in the textbook form:

1. A long black candle in a downtrend.
2. A black harami: a smaller black body that opens and closes inside the first body, but with a lower shadow reaching below the first candle's low. Sellers touched a new low and failed to hold it.
3. A small white candle that opens above the second candle's low and closes higher than its open. In Nison's description the third body sits below the second close, which keeps the whole pattern near its lows.

The resulting silhouette, with its descending first body, a lower wick reaching down like a river bed, and a small white candle resting at the bottom, is the "river".

The exact rules Finabase uses

Finabase uses TA-Lib's CDLUNIQUE3RIVER on daily bars:

Two places where TA-Lib is looser than the books:

The third candle can sit high. TA-Lib only requires the white candle to open above the second low. The source code's own comment says it is "better" if the third body is under the second close, but the test does not enforce it. In our sample only 6 of the 44 detections had a third body entirely below the second close; the other 38 had a third candle reaching at or above it.

No size test on the second body. It must fit inside the first body, which makes it smaller automatically, but it can be almost as large.

As with most TA-Lib candle functions, there is no downtrend check. 60.5% of the measurable detections came after a ten-session decline.

What the three sessions say

The first long black candle shows sellers firmly in charge. The second session opens inside that black body, a hint that the selling is not accelerating, and then breaks to a new low. That is where the pattern becomes interesting. Prices undercut the prior low, which triggers stops and draws in late sellers, and then the session recovers to close inside the first body. The new low was rejected within the day.

The third session tests whether that rejection had substance. It opens above the second-day low, trades in a narrow range, and closes higher than it opened. Neither big nor emphatic, it is simply the first session in the sequence that buyers won. Traders reading the pattern bullishly are betting that the rejected low on day two marks the bottom.

Confirmation

Because the third candle is small, the pattern needs follow-through more than most:

• A close above the third candle's high, the rule our statistics use; 18 of the 44 cases met it.
• A close above the second candle's open, which takes back the second black body.
• Support underneath: a prior low, a long-term average or a round number near the second candle's low.
• A momentum divergence, such as the Relative Strength Index (RSI): how to read and use it making a higher low while price made the new low on day two.

Two real occurrences

Intel, December 19, 2025. INTC had fallen 11.1% in ten sessions. On December 17 it printed a $1.54 black body closing at $36.05. On December 18 a smaller black body, from $37.38 to $36.28, sat inside the first one while its low of $35.67 undercut the prior low of $35.97. On December 19 a small white candle opened at $36.48 and closed at $36.82. Note that the third candle was entirely above the second close, so this is TA-Lib's version rather than the textbook one; TA-Lib also flagged a bullish harami on the same bar. The stock dipped again first, trading as low as $34.95 within the next ten sessions, below the "rejected" low, and was 0.4% lower after five sessions. Then it rallied and finished ten sessions 8.7% higher. A trader with a stop under the second low would have been shaken out before the gain.

Home Depot, March 16, 2026. HD had dropped 7.6% in ten sessions. A long black candle on March 12 closed at $338.93, a smaller black candle on March 13 dipped to $338.03 before closing at $339.03, and on March 16 a near-doji white candle opened at $342.44 and closed at $342.58. The new low on day two was only 67 cents below the first low, a marginal probe rather than a convincing rejection. The decline resumed: HD was 3.4% lower after five sessions and 5.6% lower after ten.

What the data says

CDLUNIQUE3RIVER fired 44 times on 30 of the 60 large US stocks and ETFs in our sample (daily bars, October 30, 2024 to October 29, 2026). Forty-three had a full follow-up.

The pattern performed like an ordinary day on direction and slightly worse on average return. Confirmation, which normally helps reversal patterns at least a little, left the hit rate at 50.0% and the average return slightly negative.

How it compares with similar bottoms

The closest relative is the bullish harami: the first two candles of a unique three river bottom are a harami whose second body happens to have a long lower shadow. The third candle adds a small piece of confirmation that the plain harami lacks. It is also worth comparing with the morning star, which has the same three-act structure but a much stronger third candle.

An educational outline

How practitioners typically approach it. It isn't a recommendation to trade.

1. Look at the probe. A new low that undercuts the first candle meaningfully and recovers is more convincing than a few cents' undercut like Home Depot's.
2. Prefer the textbook third candle. A small white body below the second close keeps the pattern near its lows and gives a tighter stop.
3. Wait for a close above the third high, and ideally above the second candle's open.
4. Invalidate below the second candle's low. That is the low buyers rejected; a close beneath it says the rejection failed.

How Finabase shows the unique three river

• Chart markers. Add the Unique 3 River study to any chart; Finabase marks the white third candle wherever TA-Lib finds the pattern on daily bars.
• Scanner. TA-Lib always returns +100, so a stock completing the pattern appears on the bullish list only.
• Signals tab. The Candlestick gauge on a symbol page such as AAPL's signals gives each of 61 patterns a vote from its most recent reading in the last eight bars; a unique three river is a Buy vote. The score is (buy − sell) / (buy + sell + min(neutral, max(buy, sell))), read as Strong Sell below −0.6, Sell below −0.2, Neutral below 0.2, Buy below 0.6 and Strong Buy above.

Frequently asked questions

Why is it called "unique three river"?

It is a translation of the traditional Japanese name. The pattern is sometimes listed as the unique three river bottom to stress that it forms at lows.

Is the unique three river a harami?

Its first two candles are a bearish-colored harami: a smaller black body inside a long black body. What makes it distinct is the second candle's lower low and the small white third candle.

Does the third candle have to be below the second close?

In Nison's version, yes. TA-Lib only requires it to open above the second candle's low, so Finabase's markers include many cases with a higher third candle. Only 6 of 44 detections in our sample matched the stricter version.

How rare is it?

It appeared 44 times in 29,760 daily bars of 60 large US stocks and ETFs, about once per symbol every three years on average.

Summary

The unique three river bottom tells a careful story: a long black candle, a smaller black candle that probes to a new low and recovers, and a small white candle in which buyers finally win. TA-Lib detects a looser version in which the third candle can sit well above the second close, and in two years of data on large US stocks that version performed like an ordinary day. Use the pattern to locate a rejected low, insist on a meaningful probe and a confirming close, and treat the second candle's low as the line that must hold.