Candlestick Patterns
Tweezer Top / Tweezer Bottom
Two candles print nearly the same high, or nearly the same low, and the market shows the exact same price got rejected twice.
Click two candles, test the match
Click any candle on the tape below to select it, it highlights gold. Click a second candle to test the pair, it highlights gold too. A third click clears the pair and starts a new one. The tape mixes real tweezer pairs, some adjacent, some several bars apart, with a couple of near misses that look close by eye and aren't.
Click any two candles on the tape to test them.
How it works
- The test is on the wick extreme, not the body. A tweezer top matches the two candles' highs, a tweezer bottom matches their lows. What the bodies do in between doesn't factor into the test at all.
- The two candles don't have to sit next to each other. One bar apart counts, so does several bars apart, as long as the same extreme gets tested twice.
- Tolerance is a judgment call, not a fixed law. Real markets never print two identical prices, so this drill uses a percent-of-range window instead. Tighten the slider and fewer pairs qualify.
- Opposite-colored bodies add confirmation but aren't required. A tweezer where the second candle closes opposite the first reads stronger, but the wick match is what actually defines the pattern.
- The level tested twice becomes a reference level going forward. That's exactly what the play-forward step on this page checks, whether the price that got rejected twice actually held afterward.
Where this breaks
Two candles can match by pure coincidence in a quiet market
The whole signal depends on the match meaning something, that a real level got defended twice. In a tight, low-volatility stretch, dozens of candles print highs and lows within a hair of each other simply because the session isn't moving much, not because any price level is special. The match-tightness readout on this page still lights up green in that situation. Two candles a few bars apart with highs within a few percent of the average range pass the same test whether that closeness represents genuine rejection at a real level in an active market, or plain noise in a dead one. Tolerance alone can't fix this, since tightening the slider only demands a closer number, it doesn't check whether the range around the tweezer was wide enough for a matching extreme to actually mean something. That judgment is left to the visitor.