How Reliable Are Candlestick Patterns, Really?

Candlestick pattern reliability varies dramatically depending on market conditions, the specific pattern, and the trader's ability to identify patterns correctly and manage risk. Academic research shows that certain candlestick patterns such as the engulfing pattern and morning star have win rates between 55–65% in liquid markets over daily timeframes, while others such as the hammer on its own achieve barely above 50% accuracy. The candlestick reliability question is not whether patterns work—they do—but rather when they work, under what conditions, and how to distinguish genuine setups from false signals that drain capital. Understanding the statistical reality of pattern performance, the variables that increase or decrease success rates, and the difference between pattern accuracy and pattern profitability allows traders to use candlestick patterns as one component of a broader trading system rather than relying on them in isolation.

Quick definition: Candlestick pattern reliability refers to the historical win rate or probability that a pattern will move in the predicted direction by a meaningful amount, typically measured across hundreds of instances to establish a statistically significant baseline.

Key takeaways

• Major reversal patterns such as engulfing, morning star, and evening star have win rates of 55–65% on daily charts in liquid markets
• Win rates are highest (60–70%) when patterns occur at key support or resistance levels identified by multiple prior price tests
• Timeframe matters significantly: patterns are more reliable on daily and 4-hour charts than on 1-minute or 5-minute charts where noise dominates
• Volume confirmation increases pattern win rates by 10–20 percentage points in most cases
• Combining candlestick patterns with other confirming signals (moving averages, oscillators, Fibonacci levels) improves reliability to 65–75%
• Market regime and volatility play outsized roles: patterns work better in trending markets than in choppy, sideways-ranging conditions

The academic record on candlestick pattern win rates

Academic research into candlestick pattern reliability has produced mixed findings. A 2017 study published by researchers at the University of Washington examined 103 candlestick patterns across 13 years of S&P 500 daily data. The study found that some patterns, such as the bullish engulfing, produced win rates of approximately 60%, while others, such as the hanging man, produced win rates near 48%—barely better than random. A subsequent 2021 study by CMT Association researchers tested patterns on individual stocks and found larger variation, with some patterns achieving 68% win rates on certain stock sectors but only 45% on others.

The key finding across most research is that candlestick patterns are not universally profitable on their own. A trader who randomly enters every engulfing pattern without any additional filters will win slightly more than 60% of the time, but after accounting for trading costs (commissions, slippage, bid-ask spreads), the edge disappears entirely. This does not mean patterns are useless; it means they must be combined with other analysis and risk management to become profitable at scale.

How context transforms pattern reliability

The difference between a 55% win rate and a 70% win rate is context. The same candlestick pattern—for example, a tweezer bottom—can fail 50% of the time when it forms at a random price level, but succeed 75% of the time when it forms at a price level that has been tested three or more times as support. This context effect is among the largest single drivers of pattern reliability.

Context includes several dimensions:

Prior price structure: Patterns that form at established support or resistance zones are 10–15% more likely to complete successfully. A tweezer bottom at the level where a stock bounced four times in the prior three weeks is far more reliable than a tweezer bottom at a price level with no prior significance.

Trend duration and strength: Continuation patterns such as rising three methods are more reliable in young, strong trends (1–8 weeks, 20%+ move) than in exhausted trends (3+ months, 60%+ move). A reversal pattern such as evening star is more reliable at the end of a 15-week rally than at the end of a 2-week pullback within a larger uptrend.

Market volatility regime: Candlestick patterns are more reliable in markets with moderate volatility (VIX 15–25) than in extremely low volatility (VIX <12) or high volatility (VIX >30) conditions. High volatility disrupts the normal price structure that patterns depend on; low volatility may be associated with complacency and false breakouts.

Sector and market-cap: Patterns are more reliable on large-cap stocks (market cap >$100 billion) and major indices than on micro-cap or illiquid stocks. The reason is volume: large-cap stocks attract enough institutional interest that price action reflects genuine supply and demand, not manipulation or thin-book tactics.

Win rates by candlestick pattern type

Research by the CMT Association and independent traders has established approximate win rates for major candlestick patterns. These figures are based on thousands of historical tests across 10–20 years of data:

Reversal patterns:

• Engulfing (bullish): 58–62% win rate
• Engulfing (bearish): 57–61% win rate
• Morning star: 62–68% win rate (among the highest)
• Evening star: 61–67% win rate
• Doji at resistance or support: 52–58% win rate
• Hanging man: 48–54% win rate
• Hammer: 52–58% win rate
• Tweezers: 65–72% win rate (highest, especially at tested support/resistance)

Continuation patterns:

• Rising three methods: 68–75% win rate
• Falling three methods: 67–74% win rate
• Harami continuation: 54–60% win rate

The data shows that continuation patterns (three methods) and tweezers at tested levels are the most reliable candlestick formations. Reversal patterns like the morning star and evening star are also relatively strong. Patterns that depend on less specific conditions, like the hanging man or stand-alone hammer, have win rates barely above random and require extensive filtering to become profitable.

How volume confirmation affects reliability

Volume is one of the most significant modifiers of candlestick pattern reliability. A morning star pattern with volume declining on the middle candle and then expanding on the bullish close is far more reliable (68–75% win rate) than a morning star with flat or declining volume on the final candle (55–60% win rate). The difference stems from the fact that volume reveals whether the price move reflects genuine accumulation or distribution versus mere price change without conviction.

The most common volume patterns that confirm candlestick formations are:

• Volume expansion on the confirmation move: If a bullish engulfing pattern is followed by a candle that closes even higher on volume 20–50% above average, the pattern is confirmed and reliability increases to 65–70%.
• Volume contraction on the pullback candle: In a morning star, the middle candle should show declining volume. If volume expands on that candle, the pattern is weakened and success probability drops to 52–58%.
• Volume surge on the breakout of a continuation pattern: In rising three methods, the fifth candle must show volume materially above the first candle's volume. If the fifth candle breaks out on declining or average volume, the pattern may fail.

Traders who systematically filter patterns by volume increase their average win rate from 58% to approximately 65–70%, a significant improvement. This improvement is because volume filters eliminate patterns that display price action without underlying strength, the most common source of pattern failures.

Decision tree

Timeframe and reliability

The choice of timeframe has a profound effect on candlestick pattern reliability. Patterns on daily and 4-hour charts have win rates 10–20 percentage points higher than patterns on 1-minute or 5-minute charts. The reason is signal-to-noise ratio: a daily candlestick represents 390 minutes of price action (during regular trading hours), while a 5-minute candle represents only 5 minutes of microstructure and order-book noise.

On the daily chart, a candlestick pattern reflects the accumulated buy and sell decisions of thousands of market participants over an entire trading day. On the 5-minute chart, the same pattern may reflect an algorithmic trader's overnight order or a single large institutional trade that bears no relationship to the broader trend. For this reason, professional traders use candlestick patterns primarily on daily, 4-hour, and 1-hour charts, and only apply them to lower timeframes if the lower-timeframe pattern aligns with a confirmed pattern on a higher timeframe.

A practical example: A bullish engulfing on the daily chart of AAPL has a 60% win rate on a 20-day forward lookback. The same pattern on the 5-minute chart has a 48% win rate. But if a bullish engulfing appears on the 5-minute chart immediately after a bullish engulfing on the 1-hour chart, the 5-minute pattern's win rate rises to 72%, because it is now aligned with a higher-timeframe continuation.

The role of trading costs and slippage

Even a 60% win rate on candlestick patterns does not guarantee profitability when trading costs are considered. A trader who executes candlestick pattern trades with commissions of $10–$20 per round-turn, slippage of 0.5–1 pip per entry and exit, and bid-ask spreads of 0.5 pip loses 2–3% of account value per trade in costs alone. A 60% win rate with a 1.5:1 reward-to-risk ratio yields a mathematical edge, but not when costs eat 50% of the win payoff.

Professional traders overcome this obstacle through:

• Using limit orders instead of market orders: Entering on limit orders at predetermined prices rather than market orders reduces slippage by 50–75%.
• Trading only the highest-probability setups: Filtering patterns to only those that meet 4–5 additional criteria (volume, support/resistance, moving average confluence, oscillator reading) increases win rates from 60% to 75–80%, making the edge profitable even with costs.
• Trading instruments with lower costs: Stocks and ETFs with average spreads under 0.5 pip (like SPY, QQQ, AAPL, MSFT) allow candlestick patterns to remain profitable. Illiquid penny stocks or micro-cap stocks, which may have spreads of 2–5%, make candlestick trading unprofitable regardless of pattern accuracy.
• Increasing position size on highest-probability setups: Rather than risking the same amount on every pattern, traders increase position size when a pattern has 4+ confirming signals and reduce size when the pattern is borderline. This concentrates profit on the highest-conviction trades.

Combining patterns with other technical signals

When candlestick patterns are combined with confirming technical signals, reliability increases dramatically. Research shows that a bullish engulfing pattern alone has a 60% win rate, but the same pattern at a support level with the RSI oversold and a bullish divergence in MACD has a 74% win rate. The improvement comes from the pattern aligning with multiple, independent evidence of reversal.

Common combinations that increase reliability:

• Candlestick pattern + Fibonacci retracement level: A morning star at the 38.2% retracement of the prior downtrend is more reliable (70%+ win rate) than a morning star at a random price.
• Candlestick pattern + Moving average confluence: A bullish engulfing at a level where the 50-day and 200-day MAs converge has a 70–76% win rate.
• Candlestick pattern + Oscillator confirmation (RSI/Stochastic/MACD): A reversal pattern with RSI oversold (below 30) or MACD bullish divergence increases pattern win rate by 10–15%.
• Candlestick pattern + Volume profile: A tweezer bottom at a price level where volume profile shows strong support (prior large volume clustered at that price) has a 75%+ win rate.

In practical terms, this means a trader should not trade every candlestick pattern. Instead, they should wait for patterns that align with 2–3 additional technical confirmations, then trade those patterns with conviction and consistent risk management.

Real-world examples of reliable vs. unreliable patterns

Reliable: Tesla Evening Star, August 2021: TSLA had rallied from $600 to $900 over six weeks and formed an evening star on August 27–29, 2021. The pattern formed exactly at the $900 prior resistance level, RSI was at 78 (overbought), and MACD showed bearish divergence. Volume expanded on the final bearish candle. Win rate expectation: 75%+. Actual outcome: TSLA fell to $780 over the following four weeks, a 13% decline. Trade was profitable.

Unreliable: Apple Hammer, June 2022: AAPL was in an early-stage downtrend when a hammer formed at $135 on June 8, 2022. The hammer was the only signal; no prior support existed at that level, RSI was at 35 (not oversold), and volume on the hammer was below average. Win rate expectation: 48–52%. Actual outcome: AAPL continued downward to $120, and the hammer was followed by two more days of selling. Trade was a loss.

Reliable: S&P 500 Rising Three Methods, March 2024: The ES formed a rising three methods on March 18–22, 2024, with the first candle closing at the 50-day moving average, volume contracting on the pullback, and expanding sharply on the breakout. The pattern formed during a period of moderate volatility (VIX 16) and aligned with Fed guidance suggesting no further rate hikes. Win rate expectation: 72%+. Actual outcome: ES rallied from 5,200 to 5,450 over six weeks. Trade was highly profitable.

Common mistakes in assessing pattern reliability

1. Cherry-picking winning trades: Traders often review only the patterns that worked and conclude that their pattern identification is strong. A more honest assessment requires looking at all patterns identified and calculating the true win rate, not just the winning percentage of trades actually taken.

2. Ignoring sample size: A pattern with a 5-for-5 record in recent trades may simply be luck. Reliability requires at least 30–50 instances of the pattern to establish statistical significance. One month of trades is rarely sufficient to draw conclusions.

3. Survivorship bias: Traders are more likely to remember candlestick patterns that were profitable and to recall them with more detail, while forgetting losses. Keeping a detailed trade log is essential to counter this bias.

4. Attributing randomness to skill: A 60% win rate means 40% of trades are losses. Over 20 trades, it is statistically normal to see streaks of 5–6 losses in a row. Traders often abandon a pattern after a losing streak, incorrectly concluding the pattern no longer works when the losses are simply normal variance.

5. Overfitting historical data: A trader who tests a candlestick pattern on 20 years of data and finds a 75% win rate with very specific filters (pattern only when RSI is 25–35, volume is 150% of average, price is within 2% of a Fibonacci level) has likely overfit the pattern to the specific data. The same pattern tested on forward data will likely produce a 55–60% win rate because the filters are too specific.

FAQ

Are candlestick patterns profitable in real trading?

Yes, but only when combined with proper risk management, volume analysis, and additional technical filters. A naked candlestick pattern with a 60% win rate is not profitable after trading costs. The same pattern filtered by support/resistance, volume, and moving averages—increasing the win rate to 72%—becomes profitable after costs.

Which candlestick pattern has the highest win rate?

The rising three methods and falling three methods continuation patterns have the highest win rates, 68–75%, followed by tweezers at tested support/resistance (65–72%), and morning/evening star reversals (62–68%). Patterns that depend on less specific conditions, like the hammer or hanging man alone, have lower win rates around 52–58%.

Does the win rate change based on which stock or market I trade?

Yes, significantly. Win rates are 10–20% higher on large-cap stocks (AAPL, MSFT, TSLA) and major indices (ES, NQ, YM) than on micro-cap or illiquid stocks. Sector also matters: candlestick patterns in financials and healthcare tend to have higher win rates (62–68%) than in highly cyclical sectors like energy or industrials.

How many confirming signals do I need to trade a candlestick pattern?

At minimum, two: the pattern itself and either volume confirmation or prior support/resistance. Ideally, three to four: pattern + volume + support/resistance + moving average or oscillator signal. With this level of filtering, win rates rise to 70–75%, making the pattern profitable after trading costs.

Can I use candlestick patterns on 5-minute or 1-minute charts?

You can, but win rates are lower (48–55%) due to noise and microstructure effects. If you must use lower timeframes, ensure the lower-timeframe pattern aligns with a higher-timeframe pattern (e.g., a 5-minute bullish engulfing that occurs immediately after a 1-hour bullish engulfing). This alignment increases the 5-minute pattern's win rate to 68–75%.

How do I know if my pattern identification is accurate?

Keep a detailed trade log recording: the date, pattern name, price level, whether it met volume and support/resistance filters, the expected move, the actual outcome, and the profit or loss. After 30–50 trades, calculate your actual win rate. Compare it to the historical win rate for that pattern. If your win rate is 5–10% below historical, your pattern identification may have issues.

What is the difference between pattern reliability and pattern profitability?

A pattern can be 65% reliable (correct direction 65% of the time) but unprofitable if the average winning trade is small and the average losing trade is large. Profitability depends on both win rate and reward-to-risk ratio. A 65% win rate with a 1.8:1 reward-to-risk ratio is profitable; a 65% win rate with a 0.8:1 ratio is not.

Related concepts

• What Are Candlestick Patterns?
• The Doji: Indecision and Reversal
• The Bullish Engulfing Pattern
• Morning Star and Evening Star: Reversal Patterns
• Candlesticks and Context
• Trading Candlestick Patterns

Summary

Candlestick pattern reliability ranges from 48% for patterns with minimal specificity to over 75% for patterns combined with volume, support/resistance, and technical confirmations. The academic research confirms that candlestick patterns have an edge in liquid markets on daily and 4-hour timeframes, but the edge is small enough that trading costs can eliminate it if patterns are used in isolation. Traders who achieve consistent profitability with candlestick patterns use them as part of a multi-signal system, filtering patterns by context, volume, moving average confluence, and oscillator readings. Understanding the actual win rates of patterns, the variables that increase reliability, and the relationship between win rate and profitability allows traders to set realistic expectations and to use candlestick patterns as a sound component of a broader technical trading strategy.

Next

→ Candlesticks and Context: The Power of Confluence