Single candlestick patterns: separate shape, context and confirmation
A single-candle pattern uses the body and shadows of one completed OHLC bar. A reversal label such as hammer or shooting star also needs a preceding trend. The pattern may be complete on that bar, while any rule requiring a later confirmation bar becomes available later. Neither event establishes an entry price or a profitable outcome.
Key Takeaways
- Measure shape from OHLC; record prior trend separately.
- A hammer and hanging man can share the same geometry but use opposite prior contexts.
- Body and wick tolerances vary between detectors; a name is not a numerical specification.
- OHLC cannot reveal every intrabar trade or who traded.
- A target-to-stop ratio describes planned distances, not realized returns or a guaranteed win rate.
Classify geometry before assigning an interpretation
Let the body be the absolute difference between close and open. The upper shadow is high minus the larger of open and close; the lower shadow is the smaller of open and close minus low. Their sum equals high minus low. Use ordinary price candles and state any tick tolerance. Exclude a zero range before calculating ratios.
| Shape | Visible price structure | Additional condition to state |
|---|---|---|
| Hammer / hanging man | Small body near the high, long lower shadow | Prior decline for hammer; prior advance for hanging man |
| Shooting star / inverted hammer | Small body near the low, long upper shadow | Prior advance for shooting star; prior decline for inverted hammer |
| Doji | Open and close equal or close under a stated tolerance | Specify the tolerance denominator and handling of zero-range bars |
| Full marubozu | Open and close at opposite extremes, without shadows | Specify exact equality or tick tolerance; does not prove an uninterrupted move |
| Spinning top | Small body with shadows on both sides | State body and shadow limits; equal shadow lengths are an optional restriction |
| Pin bar | Small body with one dominant shadow | Define ratios and context; overlap with other names is possible |
The same hypothetical bar can receive different labels
Consider open 100, high 101.2, low 97 and close 101. The body is 1, upper shadow 0.2, lower shadow 3 and range 4.2. A custom shape rule requiring body at most 25% of range, lower shadow at least twice the body and upper shadow at most 10% of range accepts it: the relevant ratios are 23.81% and 4.76%. These prices and limits are illustrative, not an observed trade or platform defaults.
- After a decline defined before the candidate, this geometry can be studied as a hammer candidate.
- After an advance, the same geometry can be studied as a hanging-man candidate.
- If prior prices meet neither selected trend test, retain the geometric description instead of forcing a reversal label.
- A doji threshold of body at most 5% of range would reject this same bar. Different pattern families need their own conditions.
- Changing only its high to 102 leaves valid OHLC but makes the upper shadow 20% of the range, which fails the stated 10% limit.
Keep candidate, confirmation and fill times distinct
The example shape becomes fixed when its bar closes. One possible bullish experiment requires the very next bar to close above 101.2 without trading below 97. A next close of 102 with high 102.5 and low 99 can qualify; a close of 100.5 cannot. A next bar with low 96.9 is rejected under this rule even if it closes at 102. Each is a separate hypothetical outcome, not a prediction.
- Expire the candidate when the chosen next-bar rule fails; do not wait indefinitely for a favorable price.
- Record confirmation on the next bar, not retrospectively on the candidate bar.
- In a default bar-close TradingView strategy, a market order normally fills on the next available tick, often the following bar open. It need not fill at the signal close.
- Define exit timing, order type, position size, fees and slippage separately. A pattern detector is not a complete strategy.
Distinguish a planned ratio from observed profitability
Suppose a hypothetical fill is 102, a stop trigger is 96.9 and a target is 112.2. The planned downside distance is 5.1 and the upside distance is 10.2, giving a 2:1 target-to-stop ratio. This calculation does not state which order fills first, whether either fills at its trigger price, or what the average realized win and loss will be.
- With an assumed 50% win rate, exactly +2R wins and -1R losses would have a gross arithmetic expectancy of +0.5R per trade. These are assumptions, not measured outcomes.
- Costs, gaps, partial exits and fill behavior can alter those amounts. A planned 2R target cannot ensure profitability.
- Volume above an average does not identify the traders behind it. An RSI threshold changes the sample but does not independently prove a reversal.
- Compare a filter with an unfiltered baseline using the same market, dates, execution model and costs, then evaluate on data excluded from parameter selection.
Choose the bar construction before comparing patterns
A chart timeframe, trading session and data feed determine the OHLC bars. A daily bar can have a different shape when session boundaries differ. Cryptocurrency trading schedules do not make a pattern profitable, and no timeframe is established as universally more reliable here. Synthetic chart prices, such as Heikin-Ashi, require separate handling when simulating fills.
- Record the symbol, feed, ordinary-candle chart type, interval and session.
- Count completed candidates separately from confirmations and executed trades.
- Retain failed and expired cases in the study.
- Consult each linked pattern guide for its exact shape and context rules instead of assuming all single-candle labels share a detector.
This page is for informational purposes only and does not constitute investment advice. Trading stocks, forex, and crypto carries substantial risk of loss. Past performance does not guarantee future results. Always consult a qualified financial advisor before making trading decisions.