Tweezer tops and bottoms

Tweezers are taught as a reversal pattern. The study says otherwise, and it is the sharpest contradiction in this whole file: the tweezer bottom, supposed to announce a bullish reversal, is measured as a bearish continuation 52% of the time. Not slightly worse than advertised, the opposite. This lesson explains the pattern, that number, and why it is not so surprising once you look at the mechanism.

The pattern, in two candles

Tweezer bottom, tweezer top Two candles sharing exactly the same extreme. On the left, a tweezer bottom: the same low, in a decline. On the right, a tweezer top: the same high, in a rise. The definition says nothing about body colour or size, it rests solely on the equality of the extremes, and that is what makes it rare.

A tweezer bottom is made of two candles sharing the same low, in a downward trend. Two shadows stopping at the same price, like the two arms of a pair of tweezers.

The tweezer top is its mirror: two candles sharing the same high, in an upward trend.

The definition says nothing about body colour or size. It rests only on the equality of the extremes. That is an unusually strict definition for a candlestick pattern, and it is what makes it rare.

The expected reading is intuitive: the market tested the same level twice and did not get through. A floor is forming.

The number that flips the reading

Study by Thomas Bulkowski for the tweezer bottom, quoted as it stands.

Theoretical performance: bullish reversal. Tested performance: bearish continuation 52% of the time.

That line deserves a pause, because it is unique in this file. The other patterns are measured worse than their reputation. This one is measured the other way round. The pattern supposed to signal the end of a decline most often accompanies its continuation.

Frequency rank: 39. It is rare, which is usually a good sign. Overall performance rank: 44 out of 103, so in the upper half, which is better than the hammer and far better than the engulfing.

Price target met: 71% in a bull market with an upward breakout. Best average move over ten days: 4.95% in a bear market with an upward breakout.

Source: thepatternsite.com, consulted in September 2026.

Measured the other way: 52% bearish continuation The tweezer bottom is taught as a bullish reversal and measured as a bearish continuation 52% of the time: it is the only pattern in the category measured the other way. Yet it ranks 44th out of 103 for performance, in the top half, and it is rare, with a frequency rank of 39. The pattern is not bad, its usual reading is.

Why it is not that surprising

Two candles stopping at exactly the same price do not necessarily describe a floor. They describe that there is something at that price. That something may be a buyer, in which case the classic reading holds. It may also be a resting order, or a round number, or yesterday's low that a lot of people are watching.

And in that last case, a level everybody is watching is a level where stops pile up. The market comes back for it, goes through, and continues. That is exactly the liquidity grab mechanism, and it explains a bearish continuation far better than a reversal.

The general lesson, valid beyond this pattern: a price equality is not proof of defence. It only proves the same thing happened twice in the same place.

The shared level attracts price Two candles stop at the same price: there is something at that price, but not necessarily a buyer. A level everyone watches is a level where stops pile up. The market comes back for it, goes through, and carries on. Equal prices are not proof of defence, and the shared level reads as a target rather than a support.

What you can do with it

A pattern measured against its reputation is not unusable, it simply has to be read the other way from what you were taught.

Treat the shared level as a target, not as support. Price has a good chance of going for it and through it. If you are long just above, you are in the way.

Do not enter a reversal without confirmation. Given the number, going long on a tweezer bottom means taking the minority side. You should at minimum wait for price to break the high of the two candles.

And remember the frequency rank of 39: the pattern is rare, so you will see few of them. A rare pattern going the wrong way barely more than one time in two is not a basis for a strategy, it is context.

What it gives in your own journal

The question to settle is simple and nobody asks it: for you, does a tweezer announce a reversal or a continuation?

Tag the tweezers you spot, whether or not you take a trade, and record what price did next. Twenty observations are enough to see whether the 52% shows up for you. That is an observation record, not a trading one, and it is the only way to know.

If you do take trades on them, also measure maximum adverse excursion. On a pattern whose shared level attracts price, it should be large, and it will tell you where to place a stop that survives.

Tradoshi lets you label a trade without tying it to a strategy, so the observation files in the same place as everything else.

Frequently asked questions

Must the two lows be at exactly the same price? That is what the definition says. The more tolerance you allow, the more you are counting something else, and the less the quoted figures apply.

Does candle colour matter? No, the definition does not mention it.

Why is it so rare? Because exact equality of two extremes is a strict condition. Frequency rank 39.

Should it be abandoned? No, it should be read differently: the shared level is a likely target rather than support.

And the tweezer top? It is the mirror, two identical highs in an uptrend. The same caution applies.

Key takeaways

Two candles sharing exactly the same extreme. A strict definition, hence a frequency rank of 39: the pattern is rare.
Supposed to announce a bullish reversal, it is measured as a bearish continuation 52% of the time. It is the only pattern in this file measured backwards.
Performance rank 44 out of 103, so in the upper half: the pattern is not bad, its usual reading is.
A price equality is not proof of defence. It is often a place where stops pile up, therefore a target.
In your journal, record what price does after every tweezer you spot, even without a trade. Twenty observations are enough.

Going further

These blog articles dig into this lesson's ideas, one subject per article.

Patterns in trading: spotting what makes you win or lose Supply and demand zones explained