Spinning top candlestick: indecision without prediction
A spinning top candlestick has a small real body with an upper and lower shadow longer than that body. It describes a session that closed near its open after trading on both sides. The shape is usually read as indecision; it does not establish whether the next move will reverse or continue the trend.
Key Takeaways
- Measure the body and both shadows. Two long shadows matter more than whether the candle closes green or red.
- A doji has an almost absent body. A spinning top has a small visible body; the numerical boundary is a rule you must specify.
- Classify the prior trend before seeing the following bars. A later rise or fall cannot retroactively make the original candle directional.
- Separate a completed pattern, a later confirmation condition, and a trade outcome. Each answers a different question.
Define the body and both wick ratios
TradingView describes both white and black spinning tops as small-bodied candles with upper and lower shadows longer than the body. Exact equality between the two shadows is not required by that description. For a repeatable detector, define range R = high minus low, body B = absolute difference between close and open, upper shadow U = high minus the larger of open and close, and lower shadow L = the smaller of open and close minus low. Skip R = 0 before dividing.
| Bar | Open | High | Low | Close |
|---|---|---|---|---|
| Spinning top | 100 | 103 | 97 | 101 |
| Doji | 100 | 101 | 99 | 100 |
| Long-leg doji | 100 | 104 | 96 | 100 |
Hollow bodies close at or above the open. Filled bodies close below the open.
Hypothetical prices in arbitrary units. The spinning top has R = 6, B = 1, U = 2 and L = 3: its body is 16.7% of range and both shadows pass the example rule. The other candles have zero bodies; the long-legged example has wider shadows on the same price scale. No market, timeframe or performance is represented.
- Illustrative rule: require B/R greater than 5% and no more than 25%, with U at least 1.5 times B and L at least 1.5 times B.
- Treat B/R of 5% or less as the doji bucket in this example. These thresholds are chosen for demonstration, not universal definitions or TradingView defaults.
- If wick symmetry matters to your study, add a separate maximum ratio between the longer and shorter wick. Record that filter instead of judging symmetry by eye.
Spinning top vs doji vs long-legged doji
The main distinction is the open-to-close distance. StockCharts describes doji as having virtually equal opens and closes, with tolerance depending on the instrument and surrounding candles. A long-legged doji adds long shadows on both sides. Neither that name nor a wider range establishes a higher probability of reversal.
| Feature | Spinning top | Doji | Long-legged doji |
|---|---|---|---|
| Body | Small and visible | Almost absent | Almost absent |
| Shadows | Both longer than the body | Length and balance vary | Long upper and lower shadows |
| Example detector | 5% < B/R ≤ 25%; U/B and L/B ≥ 1.5 | B/R ≤ 5% | Doji rule plus a stated shadow-length rule |
| Missing information | Next direction | Next direction | Next direction and whether the wide range persists |
Observe trend and range outcomes separately
A spinning top candlestick pattern can appear after an advance, after a decline, or within a range. Use those as separate observation groups. For example, define an uptrend in advance as a close above the 20-bar moving average with that average above its value five bars earlier; reverse those tests for a downtrend. This is an illustrative classification, not an optimal trend filter. Keep a sideways group for a separately defined range rather than calling every unclassified bar sideways.
| Context | Observation to test | What would contradict the reversal reading? |
|---|---|---|
| After an advance | Does a later close fall below the spinning top low? | Price closes above its high and continues upward. |
| After a decline | Does a later close rise above the spinning top high? | Price closes below its low and continues downward. |
| Inside a defined range | Does price leave the larger range, not just this candle? | A break of the candle remains inside the larger range or returns immediately. |
Confirmation is a later event, not a property of the candle
For an illustrative close-based rule, keep the spinning top high and low fixed after it closes. Observe the next three completed bars. A close above its high is an upward range break; a close below its low is a downward range break. If neither happens, label the setup expired. A wick through a boundary followed by a close inside does not pass this rule. Choose the window before testing and do not extend it until a preferred result appears.
- Using the chart example, a later close of 103.4 passes the upward-break test; a high of 103.4 with a close of 102.6 does not.
- If an upward break is followed by a close back inside 97 to 103, record the return into the range. The prior break occurred, but sustained upside was not established.
- An unfinished candle can change its body and shadows. Waiting for its close delays detection until the bar is complete; it does not prove the next move.
- For Pine indicators, barstate.isconfirmed can gate a chart-bar condition at the closing update. TradingView documents that it does not work inside request.security(); higher-timeframe confirmation needs separate handling.
Test a recognition rule without assuming a trading edge
Start with standard OHLC candles and a fixed symbol, timeframe, session and data source. Save the body and wick thresholds, trend label and confirmation window with each observation. Measure subsequent outcomes using a fixed horizon, such as the close five bars after the pattern, and retain expired or failed cases. The five-bar horizon is another study assumption. Comparing trend groups can show whether the same shape behaves differently in your sample; it cannot establish future performance.
- Keep the recognition label separate from any entry rule. Adding volume, RSI or support filters changes the experiment and its sample size.
- Compare with an appropriate baseline, such as other bars in the same trend group. A count of successful examples alone does not establish usefulness.
- For a strategy test, specify entry timing, exits, position sizing, fees and slippage. A confirmation close is not automatically an executable fill at that price.
- Use a later, untouched period to examine the chosen rule. Do not present a threshold selected on past results as a proven setting.
Pineify next step
Turn your chosen ratios into a reviewable script
Give Pineify AI Coding Agent the body cutoff, both wick ratios, doji exclusion and close-confirmation rule you want to test. Review the generated Pine Script and compare its labels with known candles in TradingView. If you use the Visual Editor, exposed custom-code signals can be combined with other conditions. Extra confirmation remains your choice; this page does not scan live markets or validate a profitable strategy.
This material is educational and is not investment advice. All prices and detector thresholds shown here are hypothetical. Pattern recognition does not establish a profitable strategy, and trading can result in loss.