Morning star and evening star patterns.

A morning star is three candles that can mark the end of a fall. A long bearish candle comes first, then a small candle near the low, then a long bullish candle that wins back much of the first one.

An evening star is the same shape turned upside down, at the end of a rise. In both, the small middle candle is the star, and it shows a moment when the move ran out of push.

What does a morning star look like?

Picture EUR/USD falling on a 4-hour chart. The prices below are round and illustrative, and the candlestick basics lesson explains open, high, low and close if they are new to you.

The first candle is a strong fall of 60 pips. The star dips to 1.0820 and recovers, yet it closes only 6 pips below where it opened, so neither side won that period.

The third candle rises 51 pips. To judge how much of the fall it took back, compare its close with the midpoint of the first candle's body.

That body runs from 1.0900 to 1.0840, so the midpoint is (1.0900 + 1.0840) ÷ 2 = 1.0870. The third candle closes at 1.0885, which is 15 pips above it.

Which checks make it a valid morning star?

Books and charting tools word the rules a little differently. These four checks are the part they share:

1. Price has been falling before the pattern. Three candles like these inside a sideways range have no fall to reverse.
2. The first candle has a long bearish body, longer than most recent candles.
3. The second candle has a small body, of either colour, at the bottom of the move.
4. The third candle is bullish and closes above the midpoint of the first candle's body.

Long and small have no fixed pip size, so compare each body with the last 10 to 20 candles on the same chart. Here the star's 6 pip body is a tenth of the first candle's 60.

Some third candles close just short of the midpoint. A close at 1.0865 in the example would fail check 4, and a strict reader would say no pattern formed.

Many classic descriptions assume a stock chart, where the star often gaps below the first candle. Forex trades around the clock on weekdays, so a candle usually opens at the previous close, and the gap rule is commonly dropped.

When the star is a doji, with an open and close almost equal, the pattern is called a morning doji star, and it is read the same way.

A version with gaps on both sides of the doji, called an abandoned baby, needs those gaps by definition. That makes it rare on forex charts for the same reason.

How does an evening star form at a top?

Turn the morning star over and you have an evening star. It follows a rise: a long bullish candle, a small star near the high, then a long bearish candle that closes below the first candle's midpoint.

Here is one on USD/JPY, again with illustrative prices. On a yen pair a pip is 0.01, so a move from 150.00 to 150.80 is 80 pips.

The first candle's body runs from 150.00 to 150.80, so its midpoint is (150.00 + 150.80) ÷ 2 = 150.40. The third candle closes at 150.20, 20 pips below that midpoint, so the four checks, mirrored, all pass.

The star's long upper wick tells part of the story. Buyers pushed price to 151.10 during that period, then sellers drove it back to 150.86 before the close.

The star closed 6 pips higher than it opened, and that does not change the reading. What counts is its small body at the top of the move and the strong bearish candle after it.

How could a trader plan a trade around a morning star?

A pattern is a picture of past prices. A trade built on it still needs an entry, a stop and a size, all decided before the order goes in.

The plan below uses the EUR/USD example to show the mechanics. The account, the risk rule and the target are illustrative choices for study.

1. Wait for the third candle to close. Until it closes above 1.0870, the pattern has not formed. Say the buy then opens at 1.0885, the third candle's close, with the spread left out to keep the sums simple.
2. Place the stop loss below the star's low of 1.0820, with a 5 pip buffer, at 1.0815.
3. Measure the risk: 1.0885 minus 1.0815 = 0.0070, which is 70 pips.
4. Size the trade. On a $5,000 account with a rule of risking 1% per trade, the limit is $50. Then $50 ÷ 70 pips = about $0.71 per pip.
5. On EUR/USD, 0.01 lot is worth $0.10 per pip, so 0.07 lot, at $0.70 per pip, is the largest size within the limit.
6. Set a target at 1.5 times the risk: 70 × 1.5 = 105 pips, so 1.0885 + 0.0105 = 1.0990.

If the stop is hit, the loss is 70 × $0.70 = $49.00. If the target is reached, the gain is 105 × $0.70 = $73.50. Both sums leave out the spread and any commission.

Steps 4 and 5 are ordinary position sizing, and the lot size calculator repeats them for other pairs. The 1.5 target is one choice among many, and the risk-reward lesson shows how it ties to win rate.

For the USD/JPY evening star the same plan runs the other way: a sell at the bid once the third candle closes, with the stop a few pips above the star's high of 151.10.

Waiting for the third candle has a cost. By the time it closes, EUR/USD has already climbed 65 pips from the star's low, so the stop sits far from the entry and the same $50 risk allows only a small position.

Where do these patterns fail?

The star shows a pause. In a strong trend a pause can be only that, and the next candle can break the star's low and keep falling.

Location changes how much a star says. A star at a level where price has turned before, such as old support or resistance, shows rejection at a price others watch. One in open space has less behind it.

Timeframe matters too. On a 5-minute chart the shape appears many times a day, and much of it is noise. On a daily chart it is rarer, but each candle covers a wide range, so stops grow.

News can fake the shape. A data release can print a long candle in minutes, and the next two candles may only be the market settling, with no change of direction behind them.

The evidence is weaker than the confident names suggest. Published tests of candlestick patterns give mixed results that change with the market, the period and the exact rules used, so no fixed success rate applies to your chart.

If you want to use either pattern, write your rules down exactly and test them on past data, as the backtesting lesson explains, before any real money is involved.