Matching low candlestick pattern: equal closes as support

A matching low is two consecutive down candles that close at the same price. The market falls on the first day, falls again on the second, and yet ends the second session exactly where it ended the first. To candlestick traders that repeated close looks like a floor: twice in a row, sellers drove price down to one level and could not settle it any lower.

The idea is the candle-chart version of a double bottom compressed into two sessions. A single close at a level could be chance; a second close at the same level suggests buyers are defending it. Whether that defense holds is another matter, and the data on this page gives a sober answer.

Below you will find what the pattern looks like, the surprisingly loose rule TA-Lib's CDLMATCHINGLOW applies on Finabase's charts, a successful case from Nike and a failed one from Microsoft, and a test of all 461 matching lows found on 60 large US stocks and ETFs over two years.

Anatomy of a matching low

1. A decline is under way. Prices have been falling for several sessions.
2. The first candle is black, classically a long one that extends the decline.
3. The second candle is also black and closes at the same price as the first. It usually opens higher than the first day's close, sells off during the session, and comes to rest at that same closing level.

Notice what the name refers to. It is the closes that match, not the intraday lows. Two candles whose shadows touch the same low but close at different prices are a tweezers bottom, a related but separate idea that TA-Lib does not detect as its own function.

The exact rules Finabase uses

Finabase's Matching Low chart study and scanner run TA-Lib's CDLMATCHINGLOW. Its test is short:

That is all. Several things the textbook pattern implies are missing:

No long first candle. Many descriptions call for a long black first day. TA-Lib's code does not test body size at all, so two small red candles closing at the same price qualify. This is why the pattern is so common in our data.

No trend test. The source code comment calls the pattern bullish and leaves the trend to you. Of the 451 measured cases with ten sessions of history before them, 180 came after a ten-session rise, where there is no decline to stop.

"Equal" is a band, scaled to volatility. For a stock whose recent average daily range is $4, the second close may sit up to 20 cents above or below the first. In a quiet ETF the band can be a few cents. In practice the second close can even be slightly higher than the first.

The logic of a repeated close

The close is the price that settles the day's argument: it is what funds mark their books at and what most traders see last. When two sessions in a row close at one price despite both being down days, it tells you selling pressure ran into the same wall twice. Buyers who missed the first close at that level were willing to absorb supply there again.

The bearish reading of the same two candles is simple: two red days in a row, still falling. The matching close is information about where demand showed up, not proof that it will keep showing up. That is why the pattern is better thought of as a marked level than as a signal.

What counts as confirmation

Since the matching low names a price level, confirmation comes from how price behaves around that level:

• A close above the second candle's high, the rule used in our test. Buyers took the session outright.
• A white candle on the third day that opens near the matched close and rises.
• A successful retest: price later dips back toward the level and closes above it again, a third defense of the same price.
• Oversold momentum such as an Relative Strength Index (RSI): how to read and use it reading below 30 that starts to turn up.

The reverse matters as much. A close beneath the matched level is a clear statement that the support failed, and it gives the pattern a natural invalidation point.

Two real matching lows

Nike, April 11, 2025. In the tariff sell-off of early April 2025, NKE fell 14.1% in ten sessions. On April 10 it opened at $57.23 and closed at $54.40. On April 11 it opened at $54.40, traded as low as $52.86, and closed at $54.39, a cent below the previous close and close enough to a doji that the "black" second candle is a technicality. The following session closed at $55.41, above the pattern's $55.03 high. Ten sessions later NKE was 5.4% above the matching close. A cleaner textbook case is hard to find: a sharp decline, two closes pinned to one price, and confirmation the next day.

Microsoft, June 12, 2026. MSFT had dropped 13.2% over ten sessions. It closed $390.34 on June 11 and $390.74 on June 12, 40 cents apart on a stock whose recent daily range averaged about $11, so well inside TA-Lib's band. The next session even confirmed it, closing at $399.76. Then the floor gave way: by June 17 MSFT closed at $378.91, below the matched level, and ten sessions after the pattern it was 5.7% lower. This is the case that teaches the most. Confirmation is evidence, not a guarantee, and the level that defined the pattern also defined where the idea stopped working.

What 461 matching lows did next

We applied CDLMATCHINGLOW to daily bars for 60 large US stocks and ETFs from October 30, 2024 to October 29, 2026. Because the rule is loose, the pattern is common: 461 detections across all 60 symbols, roughly one every 65 trading days per stock. We measured returns from the second candle's close.

Barely different from a random day. 54.4% higher versus 53.1% is within the noise you would expect from a few hundred cases.

The textbook setting helped a little. Matching lows after a ten-session decline were higher 55.7% of the time with a 0.54% average gain over five sessions; those after a rise did no better than chance and averaged almost nothing.

Confirmation added little. Unlike some patterns in this study, requiring a close above the pattern's high only nudged the hit rate up, to 55.8%. The ten-session average of +1.06% looks better than the baseline's +0.66%, but with returns this noisy the difference is not one to lean on.

Matching low versus its neighbors

Matching lows and homing pigeons overlap often. In our data 71 bars triggered both detectors, because a small red candle that closes a hair above the prior close can satisfy both definitions on the same day.

An educational outline for trading off the level

An illustration of how the pattern's level can structure a trade. It is not a recommendation.

1. Mark the matched close as the level. Everything else is measured from it.
2. Look for context: a decline of at least a few sessions, and preferably an older support level, a moving average or a gap nearby.
3. Wait for buyers to show up, such as a close above the second candle's high.
4. Define failure as a close below the level (or below the two candles' lows if you want more room), and size the position from that distance.

How Finabase shows matching lows

• On charts. The Matching Low study marks each second candle where TA-Lib returned +100. The NKE link at the top of the page opens a chart with it switched on.
• In the scanner. The candlestick scanner sorts every detection by the sign of the output. CDLMATCHINGLOW never returns a negative value, so matching lows always appear on the Bullish list, including the many that formed during a rally.
• In the Signals tab. The candlestick gauge on a symbol's Signals tab gives each of the 61 patterns one vote based on its latest non-zero reading in the past 8 bars. A matching low in that window is a Buy vote. The group score is (buy − sell) ÷ (buy + sell + the smaller of the neutral count and the larger side); 0.2 and above reads Buy, 0.6 and above Strong Buy.

Frequently asked questions

Do the lows or the closes have to match?

The closes. That is how both the classic definition and TA-Lib's code work. Matching intraday lows describe a tweezers bottom instead.

How close do the two closes have to be?

In TA-Lib, within 5% of the average daily high–low range of the five sessions before the first candle. On a stock that ranges $3 a day, that is 15 cents either way.

Is a matching low a buy signal?

No candle pattern is a buy signal on its own. In our test the matching low performed about like an average day. It is more useful as a defined support level to watch.

Why does Finabase show a matching low after a rally?

TA-Lib's function checks only the two candles and always reports a bullish +100. The trend before the pattern is left for the reader to judge.

Summary

A matching low is two down days that close at the same price, a hint that buyers are defending a level. TA-Lib's definition is looser than the textbook one, requiring only two red candles and nearly equal closes, which makes the pattern frequent and its average signal weak: 54.4% of 456 cases were higher a week later versus 53.1% for any day. The value lies in the level itself. It tells you exactly where the support claim is tested, and a close below it tells you the claim was wrong.