WhitmanTrading

How to Trade a Pullback

To trade a pullback, confirm a trend exists on the slower timeframe, then measure how far this instrument normally retraces. Wait for price to reach a level rather than simply pausing, name one trigger that must happen there, and set the stop beyond the point that would end the trend.

Everything difficult about a pullback is that it looks exactly like the start of a reversal, because sometimes it is one. The procedure below cannot remove that, and anything claiming to is selling hindsight.

Before you start

An established trend on the slower timeframe. Higher highs and higher lows, or the reverse. Without one there is no pullback, only a range with two sides.

A measurement of how far this instrument normally pulls back. Mark the last ten retracements in the current trend and note the range. This turns “deep” from a feeling into a number.

A level the pullback has to reach before you act. A prior swing, a moving average people watch, a value-area edge. Price pausing is not the same as price arriving.

The steps

1. Establish the trend on the slower chart

A long-horizon view showing an established directional trend.
There must be a trend first, or it is just a range. Illustrative chart - not real market data.

Higher highs and higher lows for longs, the reverse for shorts. This decides which side you are permitted to take and removes half the setups immediately.

2. Measure the normal retracement depth

A candlestick chart with previous retracements marked.
Measure how far it normally pulls back before guessing. Illustrative chart - not real market data.

Ten previous pullbacks in this trend, measured. You now have a range, and anything far outside it is a different event wearing the same name.

3. Require the pullback to reach a level

Price bars approaching a marked level.
The pullback has to arrive somewhere, not just stop. Illustrative chart - not real market data.

A prior swing low, a watched average, the edge of value. Price drifting lower and stalling in open space gives you nothing to place a stop against.

4. Name one trigger that has to happen there

A candlestick chart with a single entry condition marked.
Name the one thing that has to happen there. Illustrative chart - not real market data.

A close back above the level, a reclaim of a prior bar’s high, one observable event. Written down before price arrives, not chosen while watching.

5. Place the stop beyond the point that ends the trend

Price bars with entry and stop levels drawn as horizontal lines.
And place the stop beyond the point that ends the trend. Illustrative chart - not real market data.

Below the swing low that defines the trend structure. If price goes there, the premise from step one is gone and the trade has no reason left.

6. Check participation is fading into the pullback

A candlestick chart with a volume histogram beneath it.
Fading participation into the pullback supports the read. Illustrative chart - not real market data.

A pullback on falling volume suggests a pause. One on expanding volume suggests people are leaving, which is a different situation.

7. Size from the stop distance and take it or skip it

A candlestick chart annotated with the round-trip cost.
Every attempt costs 2% of a typical bar. Illustrative chart - not real market data.

Divide what you will risk by the distance. A deeper pullback means a wider stop and therefore a smaller position, which is the arithmetic protecting you automatically.

How to tell it worked

Score your last 20 pullback trades, taken over at least 60 days.

Count how many had a trend confirmed on the slower chart before entry. 20 out of 20. Anything less means some of those were range trades misfiled, and they will have behaved very differently.

The first half of the price series showing short directional runs.
Direction runs average 2.01 bars, so most pullbacks are shallow. Illustrative chart - not real market data.

Count how many arrived at a pre-marked level. Again 20 out of 20, or the entries were taken in open space and the stop placement was arbitrary.

Then compare your entries against your measured retracement range. Entries consistently deeper than the normal range mean you are buying reversals and calling them pullbacks.

Why depth is the whole signal

The second half of the price series with a deep retracement marked.
A deep one is often a reversal wearing a pullback's clothes. Illustrative chart - not real market data.

Measured behaviour says most moves are brief. On the shared price series, direction runs average 2.01 bars across 286 runs, with the longest at 11. Short runs mean shallow, frequent pullbacks are the normal texture of a trend.

The efficiency ratio makes the same point from another angle. Its median is 0.34 with only 30% of readings above 0.5 — most movement is back-and-forth rather than directional, so a trend is a sequence of small retracements by construction.

Which is why an unusually deep pullback deserves suspicion rather than enthusiasm. It is outside the behaviour the trend has been showing, and the most common explanation for that is that the trend has changed.

Three tests that separate a pullback from a reversal

None of the three is conclusive alone, and together they are still not certain. What they do is force the question to be answered with evidence rather than with preference, before the position exists.

Test one is depth against your measured range. A retracement inside the range you recorded in step two is ordinary behaviour for this trend. One well outside it is behaviour the trend has not shown before, and new behaviour is the definition of a change.

Test two is what happens to the structure. A pullback leaves the prior swing low intact. Once that low breaks, the sequence of higher lows has ended and the trend from step one no longer exists on the timeframe you were trading it.

Test three is participation. Falling volume into the retracement suggests absence of sellers rather than presence of them. Expanding volume on the way down means people are leaving, and that is the pattern reversals are built from.

The original data

Of the 24,971 unique videos in research/search-study-corpus.jsonl, 12 have an instruction-shaped title mentioning pullbacks, at a median of 20,253 views across 11 channels, with a maximum of 486,507. Breakouts, the opposite entry style, appear in 103 at a median of 9,904. The counts come from site/rank_howto.py.

A rising stretch of the price series cut short at the decision bar.
It is pulling back into the level. Buy? Illustrative chart - not real market data.

Twelve videos on pullbacks against 103 on breakouts, and the pullback median is twice as high. The less-covered entry style has the larger audience per video — an eight-fold supply difference in favour of the subject people are searching for less successfully.

The answer to the question on that chart is that arriving at the level is not the trigger. The trigger is the thing you named in step four, and it has not happened yet. Buying on arrival rather than on the trigger is entering before the evidence, which is the difference between a plan and a guess about where a fall stops.

When it fails

A sideways, range-bound candlestick series.
In a range every bar is a pullback from something. Illustrative chart - not real market data.

In a range the concept dissolves entirely. Every move down is a pullback from the recent high and every move up is a pullback from the recent low, so the method generates constant signals in both directions and each one is contradicted a few bars later. The word “pullback” implies a trend to pull back from, and without one the setup is just buying weakness in something going nowhere — which is why step one is not optional.

A candlestick series with several gaps, the largest marked.
And a gap turns a pullback into a new situation. Illustrative chart - not real market data.

The second failure is a gap through the level. The entry never triggers or fills far below, and the stop is already behind.

A third is entering on arrival rather than on the trigger. It removes the only evidence in the setup.

A fourth is a pullback deeper than the measured range. That is usually a reversal.

A fifth is a stop inside the pullback. The move you are trading will touch it.

And a sixth is taking one against the slow chart. The trend it is pulling back from is the wrong one.

Pullback covers what separates one from a reversal and why that is only clear afterwards. Trend following is the broader approach this entry belongs to. And support and resistance is where the levels in step three come from.

What I actually do

What fixed this for me was measuring rather than eyeballing. I had been calling pullbacks deep or shallow by feel, which meant the word changed meaning depending on how much I wanted the trade. Marking the last ten retracements on the instrument gave me an actual distribution, and after that a pullback was either inside the normal range or it was not.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.