Tristar candlestick pattern: three dojis and a turn

A tristar is three doji candles in a row, with the middle one standing apart from the other two. A doji is a session that closes almost exactly where it opened, a picture of indecision. One doji means little. Three in a row, with the middle one gapped away, suggest a market that has stopped trending and is turning in place.

The pattern comes in two directions. A bullish tristar has the middle doji below the other two and is read as a potential bottom. A bearish tristar has the middle doji above and is read as a potential top. Nison calls the formation rare and describes it as a significant reversal when it appears after a sustained move.

This page shows both versions, explains the doji test TA-Lib uses (and why it can count a candle with a body over a dollar as a doji), looks at a run of four alternating tristar signals on Texas Instruments in April 2025, and reports how 44 tristars on large US stocks behaved.

Anatomy of a tristar

The bullish version, step by step:

1. First doji after a decline. The selling pauses.
2. Second doji, gapped lower. Prices open below the first doji's body and again close almost where they opened. The decline has tried one more step but produced only indecision.
3. Third doji, higher than the second. Its body sits above the second body. The low point of the formation is behind it.

The bearish version is the mirror image: three dojis after an advance, the middle one gapped above the first, the third back below the middle.

The exact rules Finabase uses

Finabase runs TA-Lib's CDLTRISTAR on daily bars.

Three details matter in practice.

One yardstick for all three dojis. TA-Lib measures all three bodies against the average range computed before the first candle. In a calm market that threshold is small. In a violent one it can be large: before the April 2025 Texas Instruments example, TXN's average daily range was $12.14, so any body up to about $1.21 counted as a doji.

The gaps are between bodies. Shadows may overlap freely. Only the middle doji has to gap; the third only has to sit back on the other side of it.

No trend test. TA-Lib does not check whether the bullish tristar follows a decline or the bearish one follows a rise. In our sample, 45.5% of bullish tristars came after a ten-session decline and 40.0% of bearish ones after a ten-session rise. Most formed in choppy, directionless stretches.

Why three dojis can mark a turn

A trend needs one side to keep winning the close. A doji says neither side did. Three dojis in a row say the stalemate has lasted three sessions.

The gap gives the stalemate a shape. In the bullish case, the market opens lower on the middle day, which is the sellers' last attempt to extend the decline. They get the gap but not the follow-through: the close is right back at the open. On the third day the body moves back above that low doji. Sellers have spent their final push, and the balance is tilting toward buyers. The bearish tristar tells the same story upside down.

The trouble is that dojis also cluster in quiet markets that are going nowhere, and in volatile markets where TA-Lib's relative doji threshold becomes generous. In both cases three dojis can be noise rather than a turning point.

Confirmation

• Bullish: a close above the highest high of the three dojis. Bearish: a close below their lowest low.
• A real trend before the pattern. A tristar inside a range has nothing to reverse.
• A momentum extreme. An Relative Strength Index (RSI): how to read and use it that was oversold before a bullish tristar or overbought before a bearish one strengthens the reading.

In our statistics the confirmation rule was the next close beyond the third doji's high (bullish) or low (bearish). Only 5 bullish and 2 bearish tristars met it.

Real tristars: a cluster and a miss

Texas Instruments, April 2025. During the April 2025 tariff sell-off TXN's daily ranges were enormous, and the doji threshold rose with them. From April 14 to April 22 the stock printed a run of small-bodied candles: bodies of $0.06, $0.24, $0.68, $1.12, $0.15 and $0.89. TA-Lib found a tristar on four consecutive sessions: bearish on April 16, bullish on April 17, bearish on April 21 and bullish on April 22, each window of three candles overlapping the last. The labeled bullish signal on April 17 came after a 9.6% ten-session decline, and ten sessions later the stock was 10.8% higher. It is a fair result, but it sits among two bearish signals that also fired within the same week, which shows what happens when a relative doji test meets a market full of indecisive candles.

Wells Fargo, April 21, 2025. WFC printed three small bodies on April 16, 17 and 21 ($0.09, $0.33 and $0.19), the middle one gapped above the first and the third back below it: a bearish tristar after a 5.0% bounce. The rebound simply continued. The next close was $66.23, above the whole pattern, and ten sessions later the stock was 15.4% higher. After a market-wide shock, a few days of indecision were a pause in the recovery, not a top.

What the data says

Across 60 large US stocks and ETFs, two years of daily bars (October 30, 2024 to October 29, 2026), CDLTRISTAR fired 44 times on 29 symbols: 24 bullish and 20 bearish.

The bullish side looks better than a random day and the bearish side looks much worse: after a bearish tristar the average stock rose 3.34% in five sessions. Several bearish signals came during the sharp April 2025 rebound, as at Wells Fargo, which pulls the averages in one direction.

Tristar versus other doji patterns

All of these use the doji as the moment of balance. The tristar is unusual because every candle is a doji; there is no long candle to define the trend inside the pattern itself, which is one reason context matters so much.

An educational outline

This is a description of how traders often handle the pattern. It isn't a recommendation to trade.

1. Confirm the trend you expect to reverse. A bullish tristar needs a decline to end; a bearish one needs an advance.
2. Check the doji threshold. In a volatile period, compare the bodies with the stock's usual range. A $1 "doji" on a stock that normally moves $3 a day is not indecision.
3. Wait for the break. A close beyond the three-doji range is the confirmation.
4. Put the stop on the other side. The opposite extreme of the three dojis is where the turn would be wrong.

How Finabase shows the tristar

• Chart markers. The Tristar study marks the third doji. Overlapping detections in choppy periods appear as a run of consecutive markers, as on TXN in April 2025.
• Scanner. The scanner sorts by the sign of TA-Lib's output: +100 tristars appear on the bullish list and −100 ones on the bearish list.
• Signals tab. In the Candlestick gauge on a page such as AAPL's signals, each of the 61 patterns votes once, using its most recent non-zero reading in the last eight bars. Because only the latest reading counts, the alternating TXN signals would have voted with whichever came last. The gauge score is (buy − sell) / (buy + sell + min(neutral, max(buy, sell))): Strong Sell below −0.6, Sell below −0.2, Neutral below 0.2, Buy below 0.6 and Strong Buy above.

Frequently asked questions

Does the third doji have to gap as well?

Some references draw the third doji gapped away from the middle one. TA-Lib does not require that; the third body only has to sit higher (bullish) or lower (bearish) than the middle body.

What counts as a doji in Finabase?

A candle whose body is no more than 10% of the average daily high–low range of the ten candles before the first doji. That makes the test adapt to each stock's volatility.

Why can a stock show bullish and bearish tristars a day apart?

Each three-candle window is checked separately. When a stock prints many small bodies in a row, overlapping windows can satisfy the bullish and bearish rules on alternate days.

Is the tristar common?

No. It appeared 44 times in 29,760 daily bars of large US stocks over two years, and a good share of those came from a single volatile month.

Summary

The tristar packs three sessions of indecision into one pattern, with the middle doji marking the trend's last, failed step. TA-Lib defines the dojis relative to recent volatility, which is sensible but means a wild market can produce "dojis" with sizable bodies and clusters of alternating signals, as Texas Instruments did in April 2025. The 44 tristars in our sample are too few to measure an edge. Treat the pattern as a sign that a trend has stalled, and let a break of the three-doji range decide which way it resolves.