Because it is close enough to price to be touched in normal pullbacks and slow enough to show the trend. Many traders use 21 or 20 interchangeably; the exact number matters less than using the same one consistently.

How do I know it is a pullback and not a reversal?

You do not, in advance. The rules help: the trend should be intact (higher highs and lows, rising average) and the trigger needs a close above the previous high. The stop decides the rest.

What timeframe works best?

It works on any timeframe that trends. Many traders find the one-hour and four-hour charts cleaner than one and five minutes, where noise breaks the pattern often.

Pullback to the 20 EMA strategy: rules and example

The idea

Instead of buying a trend when it is already stretched, wait for it to rest. In an uptrend, price regularly pulls back toward its 20-period exponential moving average and then pushes on to a new high. The strategy buys that pause: the trend gives you the direction, the average gives you a place.

It is one of the most taught trend entries because it puts the stop close: just under the pullback low, rather than under the start of the whole move.

Watch it play out

Illustration, not a real chart.

Step by step

The rules

Setup

The 20 EMA is rising and price has made higher highs and higher lows.
Price pulls back and touches or slightly pierces the 20 EMA, without closing far below it.

Entry

Buy when a candle closes above the previous candle's high after the touch.

Stop-loss

Just under the lowest point of the pullback.

Target

2R, or the trend's last high, whichever you test better with.
Trend riders trail the stop under each new higher low.

Filters that help

Skip it if the pullback is deep enough to close well below the average: that is often the trend ending, not resting.
Skip it right before major news.

Why it can work

Trends rarely travel in a straight line. Traders who missed the first leg wait for a dip, and those who bought earlier add on weakness, so pullbacks in strong trends often find buyers near the average price of the recent past.

Waiting for a candle to close above the previous candle's high adds proof that buyers have actually returned, at the cost of a slightly worse price.

Worked example

EUR/USD is trending up on the one-hour chart. Price dips to the 20 EMA at 1.0990, the pullback low is 1.0985, and the next candle closes above the previous high at 1.1005: buy at 1.1005 with the stop at 1.0983, 22 pips away.

Risking $100, each lot loses $220 at the stop, so the size is 0.45 lots. A 2R target sits 44 pips up, at 1.1049.

Size this trade in the EUR/USD calculator

When it works, when it fails

Works best

In clean, persistent trends with orderly pullbacks: the kind where the chart looks like a staircase. Higher timeframes give fewer but more reliable trends.

Struggles

At the end of a trend and in choppy markets. The last pullback of a trend looks exactly like the earlier ones until it fails, and in a range the 'trend' is only a short swing.

Common mistakes

Buying every touch of the average without checking that a trend exists.
Putting the stop exactly on the average, where ordinary noise hits it.
Adding to a losing trade because 'it is at the average again'.

Variations

50 EMA pullback: Slower trends, deeper pullbacks: the same logic with the 50 EMA on daily charts.
Fibonacci pullback: Using a 50 to 61.8% retracement of the last leg as the buy zone instead of the average.

What the evidence says

Tested, mixed

Public tests disagree, or the edge has faded over time.

The broad idea, trading in the direction of an established trend, has strong research behind it: time series momentum (assets that rose over the past months tend to keep rising for a while) was documented across 58 futures markets over decades.

The specific pullback rule has not been studied rigorously in public. It is a common way to apply trend following, not a tested edge of its own, so treat it as something to test on your market.

Questions

Why the 20 EMA?: Because it is close enough to price to be touched in normal pullbacks and slow enough to show the trend. Many traders use 21 or 20 interchangeably; the exact number matters less than using the same one consistently.
How do I know it is a pullback and not a reversal?: You do not, in advance. The rules help: the trend should be intact (higher highs and lows, rising average) and the trigger needs a close above the previous high. The stop decides the rest.
What timeframe works best?: It works on any timeframe that trends. Many traders find the one-hour and four-hour charts cleaner than one and five minutes, where noise breaks the pattern often.