Three Line Strike Candlestick Pattern: Rules, Trading and Tested Results

The three line strike candlestick pattern is a four-bar formation: three candlesticks that move with the prevailing direction, then one large bar that opens beyond the third and wipes out the whole move by closing past the first candle's opening price. Gregory Morris, in Candlestick Charting Explained, classifies it as a continuation pattern, while other descriptions present the same shape as a reversal. That disagreement is a source of confusion around it, and a good reason to check the shape against data.

This guide gives the definition, both versions, how to identify and trade them, and what actually happened afterwards. Our test bench scanned full market history and detected 270 bearish and 291 bullish ones on hourly crypto charts alone, and compared each outcome with random bars on the same assets.

What is the three line strike candlestick pattern?

It is a four-bar structure that appears inside an established move. The first three candles form a small staircase. The fourth candle, the "strike", is a long one in the opposite direction that erases the three prior bars in one session. In the definition coded in the open-source TA-Lib library (function CDL3LINESTRIKE), the move is expected to resume after the strike: the strike is read as a sharp profit-taking pause, not as a change of direction. TA-Lib does not check what happened before the four bars; our bench adds that condition.

• Bullish version: in an uptrend, three green candlesticks with rising closes, then a red strike that opens above the third close and closes below the level where the first one started. Expected direction: up.
• Bearish version: in a downtrend, three red candlesticks with falling closes, then a green strike that opens below the third close and closes above where the first one started. Expected direction: down.

A naming trap: reversal or continuation?

The same shape is also published under swapped names: in that version, "bullish" is the sequence of three red candles followed by a large green one, read as an upward turn. Both conventions describe the same four bars; only the label and the expected direction differ. Check which one a scanner or tutorial uses before you follow it. Here, as in TA-Lib, the name gives the direction of the move before the strike, and our statistics test that direction. Since we also report how often the market went the other way, the numbers answer both readings.

Anatomy and identification of the pattern

The setup has strict rules. Each one removes look-alikes, so check them in order on your chart.

Bullish three line strike structure

1. An uptrend is in place: price before the pattern is at least one average true range above its level 10 bars earlier, in the upper half of its recent range.
2. Three green candles come next, each closing above the one before.
3. The second and third bars open inside or near the body before them: an orderly advance, without big gaps.
4. The fourth candle opens above the third close and closes below the opening price of the first one. Its body covers the three previous bodies: it is an engulfing bar over three sessions instead of one.

Bearish three line strike structure

1. A downtrend is in place: price before the pattern is at least one average true range below its level 10 bars earlier, in the lower half of its recent range.
2. Three red candles come next, each closing below the one before, each opening inside or near the body before.
3. The fourth candle opens below the third close and closes above the start of the first one, engulfing the three bodies.

These two diagrams are not drawn by hand. They show the median shape of the clean occurrences our bench found on daily charts across all markets, with the faded candles before it and the lines of the trade we test: entry beyond the four bars, stop on the other side, target at twice the amount risked.

Size and frequency

The structure is large. On hourly crypto charts, the median bearish occurrence spans 3.87% from top to bottom, or 2.16 times the average true range (ATR) of the asset; on daily charts, 5.94%, or 1.92 ATR. It is also rare: one bearish occurrence every 51,495 eligible daily bars and one bullish every 38,362. For comparison, a hammer appears about once every 552 daily bars. On a single stock, years can pass between two signals, so using it means scanning many assets or lower timeframes.

The psychology behind the strike

Take the upside version. Buyers push the market up for three strong sessions in a row. The fourth session opens higher again, then sellers take over and drive it under the level where the advance started. Three sessions of gains disappear in one session.

The continuation reading calls this a flush: short-term buyers take profits, late buyers are shaken out, and the advance resumes once that selling is absorbed. The reversal reading says a move that erases three others shows a real shift in sentiment. Both stories sound convincing after the fact. Price action alone does not tell you in advance which one is true, so we measured it on real chart data rather than on a few hand-picked examples.

Is the three line strike a reversal or continuation pattern?

By definition, a continuation one. In practice, neither reading wins on most charts.

• After the bearish version, the market kept falling a little more often than chance on crypto: 53.7% of hourly occurrences were down 10 bars later, against 50.9% of random bars on the same assets, but the average excess return over the rest of the market is not statistically significant.
• After the bullish version, it was up 10 bars later 47.4% of the time on hourly crypto charts, against 48.5% for random bars: no difference in either direction.
• On daily US stocks, the bullish version was followed by a rise 51.5% of the time against 54.3% for random days, and the bearish one by a fall 44.6% of the time against 45.5%. Neither gap is statistically significant, and stock data carries a survivorship bias we have not sized: most delisted companies are missing.

Trading it as a reversal means betting against the textbook, and our numbers do not support that bet either. In our data, the color of the strike candle told little about what came next.

How to trade the three line strike pattern

The classic trade follows the continuation reading. Our bench simulates exactly this plan on every occurrence, so the numbers on this page describe it.

Entry strategies and confirmation techniques

For a long trade, place a buy stop just above the highest point of the four candles; for a short trade, a sell stop just below the lowest point. The order only fills if price breaks out in the expected direction: that breakout is the confirmation.

It removes most occurrences. On hourly crypto charts, the entry was triggered on only 11.9% of bearish occurrences and 12.7% of the others. In 85.2% and 82.8% of cases respectively, the other side of the structure broke first and the trade was canceled before any money was at stake; the few remaining orders expired untouched after 20 bars or when the data ended. That protects you from many failures, and it also means very few trades.

A more aggressive alternative is to enter as soon as the strike has closed, in the direction of the prior move. It gives a better entry level but no confirmation, and we have not tested it.

Risk management and stop loss placement

Profit targets

We use a target at twice the distance to the stop (a 2:1 reward-to-risk ratio). On hourly crypto charts, short trades that reached it took a median of 17 bars; trades that were stopped took a median of 14. The next support or resistance level is another possible objective, which we have not tested.

Classic filters

Classic additions are a moving average to confirm the direction (for example the 200-period average), heavy volume on the strike as a sign that the counter-move exhausted itself, and a momentum indicator such as the RSI or MACD. We measured the first two below; we have not tested momentum indicators.

Advantages and limitations of the method

Advantages

• Objective criteria with clear visual cues, easy to automate in a scanner.
• Natural levels: the top and bottom of the four candles give the entry and the stop loss.
• The breakout entry screens out most occurrences before any money is at stake: only 14.0% of daily bullish setups triggered.
• On hourly crypto charts, both trades reached their target more often than the same trade on random dates, though neither gap passes a statistical test.

Limitations

• Rare: one occurrence every several tens of thousands of bars on most markets.
• Wide stop, since it covers the whole four-bar range.
• Conflicting definitions, so two scanners can label the same bars in opposite ways.
• No confirmed edge on any market or timeframe, in either version.
• Few occurrences on forex, indices, ETFs and commodities: fewer than 50 per timeframe for each version, too few to test the trade.

Quantitative data and success rates

Here is what our test bench measured. Each row compares the occurrences with the same measure on random bars of the same assets, so you can see whether it did better than chance.

Bearish version

Bullish version

How to read these tables

Compare them in pairs of columns. The first pair asks: did the market move in the expected direction over the next 10 bars more often than after a random bar on the same assets? The second pair tests the trade described above against the same entry, stop and target placed on random dates. Expectancy is the average result per trade in R, where 1 R is the amount at stake, before fees and slippage. "Occurrences" counts every detection; "2:1 trades" counts only those whose entry was triggered.

The closest call: bearish, hourly crypto

On hourly crypto charts, the bearish trade reached its target 31.3% of the time, against 24.1% for the same trade on random dates, with an expectancy of +0.09 R against −0.19 R. This difference does not pass even a simple test, let alone our correction for the thousands of tests we run: rated no clear edge, not a confirmed edge. Keep the scale in mind: 32 trades since 2017, from 270 occurrences on 206 coins. A positive average before costs on so few trades is well within what luck alone produces. The bullish trade on the same charts tells the same story (27.0% against 17.4%, +0.06 R against −0.48 R on 37 trades): rated no clear edge.

On 4-hour crypto charts the same trade rests on only 12 trades, too few to judge (8.3% against 22.2%, −0.51 R against −0.26 R): rated not enough data.

What does not hold

Nothing else does better. The bullish daily trade looks attractive at first glance, 41.9% of targets hit against 38.1% on random dates, but with 31 trades the gap in average result per trade does not pass even a simple test. The crypto daily rows are too small to judge at all: 14 bearish and 7 bullish occurrences, with 2 bearish trades and no bullish trade at all. On weekly charts, the bearish version was followed by a fall about as often as random weeks (45.9% against 44.2%, rated No clear edge), and its 2:1 trade, with 3 trades, reads Not enough data.

Context: moving average, volume and timing

These breakdowns are descriptive and not corrected for the number of tests: treat them as clues.

• 200-period moving average. On hourly crypto charts, strikes printed while the market was still above its 200-hour average were followed by a fall 60.0% of the time, against 51.1% below it. On the same charts, the short trades split between below the average (42.1% of targets hit, +0.26 R) and above it (15.4%, −0.16 R), on 19 and 13 trades: too few to read anything into.
• Volume. High volume (above 1.5 times the 20-bar average) barely changed the hourly crypto result: 53.7% against 53.8%. On 4-hour crypto charts, the long trade on high volume reached its target 66.7% of the time (+1.00 R), against 16.7% (−0.50 R) on normal volume, but on 6 and 6 trades: too few to read anything into.
• Timing. The signal acts fast: one bar later, 55.4% of hourly crypto occurrences were down, against 48.9% of random bars. After 20 bars, the gap had narrowed: 53.7% against 51.6%.

The method, from detection to the random-bar comparison, is published in our methodology.

Real examples from the charts

The gallery in this section shows cases picked by our bench (best results, textbook shapes, most recent, random draws); the written-up cases are on daily charts (crypto, sugar and corn futures), except one weekly case on a US Treasury bond ETF. They show two of the things that can happen to this trade: a winner, and cancellations followed by a turn against the pattern. The most common outcome of a triggered trade, the stop being hit (it closed more daily trades than the target, in both versions), appears in the gallery with the daily CVX card from March 1981.

A winning bullish trade: IEF, weekly, July 2011

On the weekly chart of IEF, a US Treasury bond ETF (prices adjusted for dividends), the weeks of June 6 to 20, 2011 were three green bars closing from 69.24 up to 69.85. In the week of June 27, the strike opened at 69.91 and closed at 68.12, below the first opening price of 68.90. The buy stop at the top of the four bars (69.95) filled in the week of July 18, with a stop at 68.02 and a 2:1 target at 73.81. The target was reached in the week of August 8, 2011, and IEF was 10.7% above the strike's close 10 weeks after the signal.

A canceled bullish trade: sugar futures, March 2016

From March 21 to 23, 2016, raw sugar futures printed three green bars closing from 16.29 up to 16.71. On March 24, the strike opened at 16.72 and closed at 15.87, below the first open of 15.90. On March 28, sugar traded under the bottom of the structure (15.76) before reaching the buy stop at 16.75: canceled, no trade. It was 7.4% lower 10 days later. The opposite reading was right here, and the breakout rule kept the trader out.

Two bearish occurrences, both canceled before a rise

On April 30, 2019, LTC/USDT completed a downside version: three red candles from April 27 to 29, closing from 72.23 down to 67.87, then a green strike from 67.86 to 74.32, above the 73.83 level where the first one started. The next day the top of the structure (74.74) broke before the bottom (66.50), so the trade was canceled. LTC kept rising and was 3.8% above the strike's close 10 days later.

On August 31, 2017, corn futures printed the same structure: three red sessions closing from 336 down to 329.5, then a strike from 329.25 to 342.25, above the first open of 338.75. The next day the top (342.75) broke: canceled again. Here too the strike started a rise, and corn was 3.7% higher 10 sessions later. Cancellation, not a trade, is the most frequent outcome: on daily charts, only 15.8% of bearish setups triggered an entry. The 2:1 trade table counts 24 trades for that row, fewer than the setups that triggered, because it leaves out trades still open when the data ended and trades excluded by the trade rules.

Comparison with similar candlestick patterns

The first three bars of the bullish version look like three white soldiers, and those of the bearish version like three black crows. The difference is context and the fourth candle: three white soldiers is read as a turn after a decline and needs no fourth candle, while the strike version comes inside an advance and adds a fourth one that erases it. Neither three-candle setup beat chance on daily charts in our tests: the three white soldiers trade reached its 2:1 target 34.4% of the time against 33.3% on random dates, and three black crows 22.1% against 26.0%.

The strike candle itself is an outsized engulfing pattern: instead of covering one body, it covers three, which makes it a stronger engulfing move on paper. The bullish engulfing candle is far more common (59,948 daily occurrences in our data) and showed no proven edge either: its daily 2:1 trade reads No clear edge. The rising three methods is the mirror logic: a long candlestick, a small counter-move, then a new push that resumes the advance. The advance block also starts with three green bars in an uptrend, but with shrinking bodies or long upper wicks, and it is read as a bearish warning.

Tools, indicators and code for detection

Charting platforms can scan for this setup, either with built-in candlestick recognition or with a community script. On TradingView, several free scripts flag it and can send alerts in real-time. The free TA-Lib library, also used in Python scanners, provides CDL3LINESTRIKE (our rule finds 86% of the bullish cases it flags and all the bearish ones); ProRealTime and MetaTrader users can find indicator code on their community forums.

Before relying on a tool, verify two things in its code: which convention it uses for "bullish" and "bearish", and how strict it is (must the fourth candle open beyond the third one, is a prior move required). Different criteria give different detections and different outcomes. A simple manual review is free and valuable: on a demo account, mark the last 30 occurrences on your market, write down the entry, stop and target you would have used, and record the outcome. It is the same test we run at scale.

Frequently asked questions

Summary of key points

• Three bars in the direction of the move, then one candle that opens beyond the third and closes past the first candle's opening price.
• TA-Lib reads it as a continuation; some websites read the same shape the other way, with swapped names.
• With a breakout entry, more than four signals in five were canceled before entry.
• Bearish version, hourly crypto: 2:1 target hit 31.3% of the time against 24.1% on random dates, +0.09 R before costs, rated No clear edge: on 32 trades, its average result per trade does not pass even a simple test.
• Bullish version and daily charts: no confirmed difference from random bars.

Final thoughts on risk management

Treat the three line strike as a precise location for a trade, not as a reason for one. Choose your naming convention, wait for the breakout entry, put the stop on the other side and size the position on that distance. Remember that of the daily short trades that hit their stop, 21.1% had first reached +1 R (the Size and trade path block of the bearish results): a winner can still become a loser.

Disclaimer: the information on this page is for educational purposes only and is not investment advice. Past performance is not indicative of future results. Trading financial markets involves the possibility of loss; consult a licensed financial adviser before you invest.

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. Statistics on this page were generated on November 5, 2026. The full methodology explains every step.