Tweezer bottom candlestick pattern and tweezer tops: price tolerance rules

Tweezer bottom versus tweezer top

The two patterns share a price-matching test with opposite directions. TradingView describes a long red candle followed by a green candle for a bottom, and a long green candle followed by a red candle for a top. A detector that allows any candle colors uses a broader definition and can return different results.

Define nearly equal with ticks or ATR

Choose one method and apply it to both directions. For tick tolerance, multiply a nonnegative whole number of ticks by the symbol minimum price increment. For ATR tolerance, multiply a chosen fraction by a previously completed ATR value. ATR measures volatility, not reversal probability. The settings below are illustrative custom rules; they are not asserted TradingView defaults.

• For tick-based price data, compare integer tick distances so floating-point rounding does not reject the inclusive boundary.
• For the ATR variant here, use a 14-bar Wilder-smoothed ATR ending on the bar before the pair. Freeze that value for the setup so the pattern candles cannot widen their own tolerance.
• Convert the ATR tolerance into whole ticks by rounding down. For example, 0.025 with a 0.01 tick permits two ticks. A tolerance smaller than one tick permits exact matches only.
• If the required ATR history is unavailable, skip the ATR test. Do not silently substitute zero or a different rule.
• A larger tolerance admits more pairs; it does not establish that their trading outcomes are better. Do not switch methods after seeing whether a trade won.

Tweezer bottom example: two ticks apart

Assume the preceding trend and first-body-size filters pass. B1 has a low of 100.00 and B2 a low of 100.02. Their difference is two ticks on a 0.01 grid, so the inclusive two-tick rule passes. An exact-match detector rejects the same pair. B2 does not need to close at the first low, and its lower wick need not have the same length as B1.

Tweezer top example: the same tolerance at the highs

Assume a qualifying prior rise. T1 reaches 110.00 and T2 reaches 109.98, again two ticks apart. The fixed-tick rule passes without changing any tolerance parameter. If T2 instead reached 109.97, its three-tick difference would fail that rule but pass the separate 0.04 ATR rule above. This boundary explains why two detectors can disagree on a tweezer top candlestick pattern.

Measure trend before the pair

Similar extremes inside a range do not establish a reversal setup. TradingView offers SMA50, SMA50 with SMA200, and no trend detection. A custom detector should state both its trend calculation and the bar on which it is evaluated. The following is one testable alternative, not a claim to reproduce the built-in indicator.

• On the completed bar immediately before the pair, require close below its 20-bar SMA and that SMA below its value five bars earlier for a bottom. Reverse both comparisons for a top.
• Define a long first candle as a real body larger than the mean real-body size of the 20 bars before the pair. Body size is the absolute difference between open and close.
• Require the opposite-color sequence in the comparison table. Reject doji candles under this version rather than assigning them an arbitrary direction.
• Skip the setup when there is insufficient history for the selected filters. Keep filter parameters fixed across bullish and bearish cases.

Separate the completed pattern from confirmation and failure

The earliest completed pattern is known at the second candle close. A later confirmation rule adds delay. For an illustrative bottom setup, wait up to three subsequent bars for a close above the higher high of the pair. Cancel first if any of those bars trades at or below one tick beneath the lower low. For a top, require a close below the lower low and cancel at or above one tick beyond the higher high. Expire either setup after the third bar if it has not confirmed.

• For B1/B2, confirmation requires a later close above 103.00; the invalidation threshold is 99.99. For T1/T2, confirmation requires a close below 107.00; the invalidation threshold is 110.01.
• If one bar reaches invalidation and closes beyond confirmation, cancel the setup under this example. OHLC bars alone do not establish a favorable order of intrabar events.
• Keep the formation label on candle two and put a separate confirmation label on the later candle. Do not backdate the later information to make the signal appear earlier.
• High, low and close can change while a candle is open. Use closed-bar conditions for final labels and alerts when comparing them with historical signals.
• Confirmation is a research condition, not an order fill. A trading test also needs explicit entry and exit rules, fees, spread and slippage assumptions.

Specify the detector before asking for a script

A Pineify AI Coding Agent request can ask for a single detector with shared tick or ATR tolerance, mirrored top/bottom rules and separate formation and confirmation labels. Include the two examples above and the three-tick rejection case as acceptance checks. Review the generated logic and compare its output in TradingView before relying on alerts.

• Record the symbol, timeframe, session and data source. Different bar boundaries can change the highs and lows being compared.
• Use ordinary price candles for these examples. Heikin Ashi and Renko supply synthetic chart prices, so matching their extremes answers a different question.
• An engulfing pattern compares real bodies; a tweezer compares extremes. One pair can qualify as both, but neither condition substitutes for the other.
• Treat RSI, volume or support filters as additional hypotheses to test. This guide establishes no universal win rate or best timeframe.