Mechanism and definition of a candlestick reversal
A candlestick reversal is a chart idea that uses one or more candlesticks to suggest a shift from one short-term price behavior to another (for example, from downward pressure to upward pressure). The assessment is based on the shape of candles (open, high, low, close) within a chosen timeframe, not on a fixed, universal outcome.
Before looking for a reversal, state your assumptions:
- What instrument you mean (the asset or FX pair) and its trading session context.
- What timeframe you use (e.g., 1H vs 15M), because candle geometry changes with aggregation.
- Whether your evaluation is purely visual on historical candles or whether it includes trading frictions (spreads, commissions, slippage).
What data you need: inputs, provenance, timeliness, and context
To assess a candlestick reversal, collect data in four categories: inputs, provenance, timeliness, and context.
1) Core price inputs (the candle data)
You need at least OHLC candle values for the full window you analyze:
- Open, high, low, close per candle.
- The candle timestamps that define the start/end of each period.
If the method you use distinguishes the “body” and “wicks,” ensure your dataset preserves the exact high/low extremes for each candle; otherwise, the candle shape can change.
2) Provenance details (where the data comes from)
Record the data source and the rules used to build candles. For example:
- Data provider or platform (and whether it is derived from bid/ask mid, last price, or another feed).
- Candle construction settings (timezone, session hours, how holidays/weekend gaps are handled, and whether missing candles are forward-filled).
This matters because the same apparent candle pattern can differ when the underlying price definition differs.
3) Timeliness and scope (what time window and how recent)
State the analysis time window and ensure that all candles come from the same extraction method. Clarify whether:
- The dataset is fully historical and fixed, or updated continuously.
- You are evaluating a single past interval or scanning many intervals.
Timeliness also includes versioning: if you later re-download data, candle boundaries and values may change if the provider backfills or corrects historical feeds.
4) Market and chart context (what surrounds the candles)
A candlestick reversal concept is rarely isolated to one candle. You need context that helps interpret whether the market is already “stretched” or whether the move is just noise. Useful context data includes:
- Recent swing structure on the same timeframe (for example, recent lower lows/higher highs).
- Volatility regime indicators derived from the same OHLC data (only as computed features from your chosen dataset).
- Volume data if your data source provides it, noting that volume definitions may vary by venue.
If you compare results across timeframes, keep the instrument and candle settings consistent, otherwise you may measure artifacts rather than behavior.
Evidence and examples of what to check (without assuming certainty)
A practical checklist for evidence is to verify that your reversal assessment is not caused by mismatched inputs:
- Reconstruct the candle geometry from OHLC values and confirm the body/wick proportions match what the chart shows.
- Verify the candle timestamps and timezone so that “the reversal candle” truly corresponds to the intended period.
- Check data completeness: ensure no missing candles in the analysis window.
- Run a sensitivity check by slightly expanding the window (e.g., include the preceding candle). If the interpretation flips purely due to one missing candle, the conclusion is fragile.
Instead of treating any candlestick shape as a standalone prediction, treat it as a hypothesis about short-term behavior that still depends on context and execution assumptions.
Limitations, risks, and failure modes to expect
At least one material failure mode is common: noise and regime change. Candles reflect aggregated price over a timeframe, so reversals may appear during random fluctuations, especially when volatility is high or spreads/fees meaningfully affect realized outcomes.
Other important limitations:
- Overfitting to a timeframe or dataset: a pattern that “worked” in one historical sample may not behave similarly elsewhere. - Data-definition mismatch: differences in price feeds or bid/ask/mid conventions can alter candle highs/lows and thus the pattern. - Missing or corrected history: updated datasets can change historical candles and timestamps.