Pullbacks that take fewer bars than the move they retrace

[Fairman]

Pullbacks that take fewer bars than the move they retrace

One of the more useful things I look at on M5 has nothing to do with price levels. It's the number of bars. Count how many candles the impulse took, then count how many the pullback is taking.

In a healthy trend the pullback is usually slower than the push. Say price drives up over six bars and then drifts back over ten or twelve. The selling is half hearted, candles overlap, bodies are small. That kind of slow retrace is where I like to look for continuation entries, because the side that made the impulse is still in control and the other side isn't committing.

When the pullback is fast, the picture changes. If the push took six bars and price gives back most of it in two or three, someone is selling with intent. That isn't a dip to buy. Often it's the first sign the move has finished, or at least that the next leg will need a lot more work.

Speed and depth together tell you more than either one alone. A deep but slow pullback, into the 60 or 70 percent area over many bars, can still be fine. A shallow but very fast one, a 38 percent drop in a single big candle, makes me careful even though the depth looks textbook.

None of this needs an indicator. You just count candles. I mark the start and end of the impulse, note the bar count, and keep an eye on how quickly the retrace is eating into it. If the retrace is already matching the impulse bar for bar, I step aside and let the chart decide what it wants to do next.

[Fairman]

It helps to look at the candle bodies during the pullback as well as the count. A slow pullback made of small overlapping candles with wicks on both sides is a market that's resting. A slow pullback made of steady full bodied candles all closing near their lows is a different animal, even if it takes the same number of bars. The second kind is an orderly sell program, and it tends to keep going. So the bar count is the first filter, and the shape of the candles inside it is the second. On M1 the same reading works, but the counts get noisy, so I prefer to judge it on M5 and only time the entry lower down.

[Fairman]

Session timing changes what counts as fast. During the London open a two bar retrace might be normal because everything moves quickly, while the same speed at 13:00 London time means a lot more. I compare against the speed of recent moves in the same session rather than some fixed number. On a quiet afternoon, even a modest burst against the trend stands out, and it often marks the point where the morning move has been fully absorbed and the market is ready to turn or flatten out. Writing the bar counts next to recent swings for a week or two gives a feel for what normal looks like on your pair and in your hours, and after that the fast ones jump out without any counting at all.

[Fairman]

News can make the whole comparison meaningless. A pullback driven by a data release is fast because of the number, not because the trend has failed, and price often resumes once the dust settles. I exclude any impulse or pullback that contains a scheduled release when I'm judging speed. Otherwise the rule tells you the trend is over every time a minor number comes out, and you end up stepping aside from perfectly good continuation setups for no real reason. A simple fix is to wait until a few bars after the release before judging the speed of anything, and to measure the impulse only from where the post news candle closed.

[FXS Prop Desk]

Fairman, counting bars in the pullback compared with the impulse is a nice way to see the trend's strength without indicators. A related measure is the volume or range of the pullback bars: a slow pullback with small bars suggests a lack of interest in the opposite direction, which often precedes a continuation. It also helps to check whether the pullback holds the upper half of the impulse leg.

[Fairman]

For entries after a slow pullback, I wait for the first candle that closes back in the trend direction and enter on a break of its high, with the stop under the pullback low. That keeps the trade small and defined. If the next bar fails and the retrace speeds up instead, the stop is close and the idea is clearly wrong. It also means I'm never buying while the retrace is still running, which is the part of the pullback that catches most people out. If the first trend candle never comes and price just goes sideways at the low of the pullback, I leave it. Sideways at that point often turns into a range for the rest of the session, and a range is not the trade I planned for.