XAUUSD: micro pullback depth that is enough

[LondonScalper]

XAUUSD: how deep a micro pullback is enough

On gold I used to enter the first one-tick pause and call it a pullback. That is how you buy slippage for a living while telling yourself you are patient. Now I define micro-pullback depth in advance relative to the impulse that preceded it, so the trigger is measurable.

Desk rules that reduced noise entries:

• Pullback must travel a minimum fraction of the prior impulse — written, not felt in the moment.
• If spread is elevated, required depth increases or the trade is skipped entirely.
• Time-stop still applies — deep pullbacks can become reversals while you negotiate with yourself.

Structure first, trigger second. Depth rules will not save a bad bias, but they stop a lot of impatient gold clicks.

How do you define “enough” pullback depth on XAUUSD without turning it into curve-fit nonsense?

On fast gold days I would rather miss a shallow pullback than pay for five shallow fakes. Depth rules feel slow in the moment and cheap in the monthly cost column.

[Shadow Trader]

I use a fraction of the impulse too, and the way I kept it from turning into curve-fitting was to allow myself exactly two numbers and never touch them. Minimum depth is 30% of the impulse on M1, and anything deeper than 62% is treated as a possible reversal rather than a better price.

The attached illustration is the kind of morning that made me write it down. Gold ran from 4,182.0 at 09:52 to 4,194.0 at 10:03, a 12 dollar impulse, so 30% puts the line at 4,190.4 and 62% at 4,186.6.

img_12573_r1.png (86.47 KiB) Viewed 20 times

The first pause at 10:05 only came back to 4,193.2, under 7% of the move. That's exactly the one-tick pullback you described, and it rolled over within two minutes.

The limit at 4,190.4 filled at 10:11, the low was 4,190.2 a minute later, and the stop sat under the 62% line. On spread I'd disagree slightly, because I don't deepen the requirement, I just skip when gold is wider than 0.35.

Where the fraction breaks down is after data, when a 30 dollar spike makes 30% a 9 dollar retrace that is really a new move. Do you cap the impulse size you'll measure from, or skip those spikes entirely?

[Fairman]

Pullback must travel a minimum fraction of the prior impulse

What fraction do you use, and measured how? On gold that choice matters a lot. A 25% retrace of a 6 dollar impulse is 1.50, which can be inside the spread plus a couple of normal ticks on a busy morning, while 50% of the same move is 3 dollars and starts to look like a reversal. Measuring the impulse by wick or by close also changes the number. One way to avoid curve-fitting it: pick two or three round values, run all of them on the same month of trades, and only commit to one if the results clearly differ. Did you do something like that, or did the fraction come from looking at charts?

[Shadow Trader]

Fairman, my fractions are 30% minimum and 62% max for a continuation pullback, measured on M1 from impulse close to close, not wicks. Wicks on gold are mostly noise on my feed, and close to close kept the same two numbers stable across three months of tagged trades.

On a 6 dollar impulse that means 1.80 to enter and 3.72 as the reversal line. Anything shallower than 1.80 is the one tick pause you described, and I skip it even when it feels patient.

I did run 25%, 30%, and 40% on the same month before locking 30%. The 25% line sat inside spread plus two ticks too often, and 40% cut my sample in half without lifting expectancy, so 30% was the clear middle.

Spread still skips the trade above 0.35 rather than deepening the fraction. Do you measure the impulse from the breakout bar close or from the swing extreme, and did a side by side month change your pick?