How to Trade a Trend
To trade a trend, define what counts as one in writing, enter on pullbacks rather than on extensions, and exit on a mechanical rule rather than by judging when the move has finished. Trend methods give back part of every winner by design.
Trend trading is entering in the direction of a sustained move and holding while it continues. The mechanics are undemanding. What makes it hard is the shape of its results: long flat stretches, frequent small losses, and a giveback at the end of every winner.
Before you start
A written definition of what counts as a trend, because the word does most of the work. A sequence of higher highs and higher lows, or price above a stated moving average. Something checkable.
An entry rule for pullbacks, since chasing extensions is where the losses are. Buying the bar that just ran is buying the least favourable price in the move.
An exit rule that does not require you to call the top. A trailing stop, a structural level, a moving average cross. Anything mechanical.
The steps
1. Write the definition down
“Higher highs and higher lows on the daily” is checkable. “Looks strong” is not, and the second one will find a trend in every market including the ones going nowhere.
2. Wait for a pullback
On this site’s shared series 95% of bars sat below a prior peak and the median drawdown was 1.36%. Pullbacks are the normal state, so waiting for one costs less patience than it sounds like.
3. Enter on evidence the pullback has ended
A close back above a level, a higher low forming. Something that distinguishes a pullback ending from a pullback continuing, because those look the same in the middle.
4. Put the stop below the pullback’s low
If price goes below the low you just entered above, the sequence of higher lows is broken. That distance sets the size.
5. Use a mechanical exit
On this site’s shared series a 1-ATR trailing stop was hit within a median of 3 bars across 562 trials and a 4-ATR one within 32. Wider trails hold longer and give back more.
6. Accept the giveback
A trailing exit fires after price has already turned, so part of the gain is always returned. Tightening the trail to avoid it converts winners into small ones, which is the more expensive error.
7. Expect long stretches of nothing
Trends are a minority of market conditions. A method that only works in them is inactive most of the time, and the temptation to trade the rest is what dismantles it.
How to tell it worked
The trend definition is written, and you can apply it in under 1 minute.
0 entries were taken on an extended bar, all of them after a pullback.
Every exit came from the rule, so 0 were taken on a judgement about the top.
And the method was inactive during non-trending stretches, rather than adapted to them.
Why the giveback is unavoidable
Any mechanical exit fires after the turn. It has to — it is defined by price moving against you, and that movement is the giveback.
The only way to avoid it is to predict the top, which converts a mechanical method into a forecasting one. That trade is available to anybody and almost nobody wins it.
The shape of the results
A minority of trades produce nearly all of the profit. Most entries are stopped out for small losses when the pullback keeps going, which is the ordinary case rather than a malfunction.
Which makes skipping trades expensive. The large winners look exactly like the small losers at entry — on this site’s shared series direction runs average 2.01 bars, so nothing at the start distinguishes a run that lasts from one that does not.
And it makes the equity curve uncomfortable. Long flat stretches punctuated by a few large moves is what the method produces, and abandoning it during the flat stretch is the most common way it fails.
Choosing the trail width
A tight trail exits early and often. It preserves more of each winner that survives and converts many winners into small ones, because ordinary pullbacks reach it.
A wide trail holds through pullbacks and gives back more. On this site’s shared series a 4-ATR trailing stop survived a median of 32 bars against 3 bars for a 1-ATR one, across 562 trials.
The right width depends on how long your trends actually last. If the moves you catch run for twenty bars, a stop that exits at three is not a risk control, it is a different strategy.
Pick one and leave it. Widening the trail after being stopped out of a move that continued, then tightening it after giving one back, produces a setting fitted to the last two trades — which is exactly the mechanism the mechanical exit was adopted to avoid.
The original data
Of the 24,971 unique videos in research/search-study-corpus.jsonl, 47 mention trend trading in the
title, at a median of 17,853 views across 46 channels, and 53% of those titles are instruction-shaped.
Trend following appears in 116 at 3,608 and range trading in 7 at 32,401. The counts come from
site/corpus_count.py.
47 videos at 17,853 against 116 on the systematic version at 3,608. A third of the coverage and five times the audience — the discretionary framing reaches far more people than the mechanical one, which is worth knowing given that the mechanical one is what actually gets tested.
The answer to the question on that chart is that extended is not a measurement. A trend method’s profits come from the moves that go further than looked reasonable — and exiting because the move has been large is exactly the decision the mechanical exit exists to prevent.
When it fails
The failure is trading through the flat stretches, and it converts a working method into a losing one. Months pass with no qualifying setup. The definition gets loosened slightly, because surely that counts. Then entries are taken on extensions rather than pullbacks, because waiting has produced nothing. By the time a real trend arrives, the account has been ground down by trades the original rules would never have permitted, and the method that would have caught it is no longer being run.
The second failure is no written definition. Every chart then has a trend.
A third is entering on extensions. That is the worst price in the move.
A fourth is exiting on judgement. It is a forecasting skill, not a rule.
A fifth is tightening the trail after a giveback. Winners become small.
And a sixth is skipping setups that look weak. They are indistinguishable at entry.
Related
Trend following is the systematic version and its measured behaviour. Trend analysis covers identifying one. And pullback is the entry this method depends on.
Giving back part of every winner was the hardest thing to accept. A trailing exit always leaves money on the table, and every time it feels avoidable. The alternative is deciding when the move is finished, which I am not good at and neither is anybody else — so the giveback is the fee for not having to.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.