Forex Chart Patterns

The same shapes on a different market

A chart pattern is a shape in a series of prices. Examples are a double top, a triangle, a flag or a head and shoulders. The general ideas are on chart patterns, and one full example is on head and shoulders. This page asks a narrower question: what changes when the prices are a currency pair?

A currency pair is a price of one currency in terms of another. If you are new to the market itself, start with forex trading for beginners. A pip is the usual small unit of price for a pair: 0.0001 for most pairs and 0.01 for pairs that include the Japanese yen. To turn pips into money for a given position size, use the pip value calculator.

What is different about forex data

It trades around the clock

Currency trading follows the sun from one financial centre to the next. On weekdays there is almost always a market open somewhere. A chart of a stock has a clear open and close each day. A currency chart has neither. Where one daily bar ends is a choice made by the platform, so a daily pattern can look different on two platforms that cut the day at different hours.

Activity changes with the session

Trading is busier when large centres overlap and quieter when few are open. That changes how a pattern forms. A peak made during a quiet hour may be a few orders moving a thin market, not a level many people cared about. A swing made during a busy hour has more trades behind it. A chart does not show which was which.

There is no central volume

Stocks on one exchange have a volume figure for every bar. Currencies mostly trade between banks and brokers, with no single place that adds it all up. The volume shown on a forex chart is usually a count of price changes (ticks) from one data source, not the amount traded. Patterns that lean on volume, such as a breakout confirmed by a surge in volume, cannot be read the same way. Tick counts are a rough stand-in at best, and they differ between sources.

Gaps are rare, except over the weekend

Because the market runs continuously, one bar usually opens close to where the last one closed. The exception is the weekend. If news arrives while the market is shut, the first price on Monday can be well away from Friday's close. A made-up example: a buy stop sits at 1.1049, and after a weekend the first price available to buy is 1.1062. The order fills about 13.0 pips worse than the level it was placed at. That is slippage, and a pattern held over a weekend can meet it.

The price you see is not the price you get

A chart usually draws the bid price. You sell at the bid and buy at the ask, which is higher by the spread. The spread is not fixed. It is often tight in busy hours and wider in quiet ones and around news. Brokers differ, so check how yours quotes it.

A worked example with made-up numbers

Take a made-up pair on hourly bars, drawn at bid prices. Price rises to 1.1050 on bar 6, dips to 1.1000 on bar 11, and rises again to 1.1048 on bar 16. That is a double top with a neckline at 1.1000. The first close below the neckline is 1.0997 on bar 19. A short trade sells there at the bid. All the numbers are made up.

The common measured-move rule takes the pattern height, 50.0 pips, and projects it down from the neckline. That gives a target of 1.0950. A person might also treat a price above peak 2 as the point where the shape has failed. The table shows the trade with a tight spread and with a wide one. A short trade is closed by buying back at the ask, so the spread comes off the move at the exit.

With no spread the move from the entry to the target is 47.0 pips. With a 1.2 pip spread, 45.8 pips are left. With a 3.0 pip spread, 44.0 pips are left. The stop works the other way: a loss at the stop is 52.2 pips with the tight spread and 54.0 with the wide one. The ratio of the net move to the loss goes from 0.88 to 0.81. Same chart, same pattern, a different result, with nothing changed but the spread. Calculated by us on 31 Oct 2026, exact. Slippage is left out to keep the count clear.

How sessions and spread change entries

Patterns are usually taken from closes, and the close of an hour is the same kind of number wherever the hour falls. What it means is not the same. Three effects matter most.

• A thin-hour break can be a spike. In a quiet hour a few orders can push price through a neckline for a bar or two, and then it returns. A rule that enters on the first close below the line will enter on these. See false breakout.
• The spread can be widest when patterns break. News often breaks a level, and the spread often widens at the same time. The entry then pays more than the same entry in a calm hour, and a stop placed just beyond a peak can be touched by the ask price while the bid chart shows nothing.
• Small patterns cost more. The smaller the target, the larger the share the spread takes. In the example the wide spread takes 6.4 percent of a 47.0 pip move. On a pattern with a 10 pip move, the same spread would take 30.0 percent. Short time frames make small patterns, so costs weigh most there.

A rule can allow for these. It can skip entries when the spread is above a limit, skip certain hours, or wait for a second close. Each of those is a setting, and each setting is another thing to test, not a fix.

How a bot tests forex chart patterns

A bot, or a forex robot, cannot use a drawing. It needs a definition that gives the same answer every time. The full steps are in how to backtest. For patterns on currency pairs, these points matter most.

• Write the definition first. How many bars either side make a swing, how close the two peaks must be, what counts as the break, how long the pattern may take. Fix these before looking at results.
• Confirm swings late. A swing high over 5 bars either side is only known 5 bars after it happens. A test that marks it on the bar it happened has looked ahead. See look-ahead bias.
• Use bid and ask. Test with real spread data if the data source has it, or with a deliberately wide fixed spread if not. Run it twice, tight and wide, and see how much of the result survives. In the example the ratio fell from 0.88 to 0.81.
• Add swap and weekend holds. A trade held over the rollover time can pay or earn a swap charge, and a trade held over the weekend can meet a gap. Check how your broker handles both.
• Test across sessions and pairs. Split the results by the hour the trade opened. A rule that works only in one session, on one pair, in one year, is more likely a fit to that sample. See overfitting.
• Count the trades. A strict pattern definition may match only a handful of chart segments. Too few trades say little about anything.
• Keep data back. Tune on one part of the data and test once on the part you kept back.

What this does not tell you

None of this says a pattern works. A pattern is a way to describe a chart, and a test only says how a fixed rule behaved on a past sample, after costs. A good past result can fail later. This page is education, not a signal.