Recognizing gap fill probability on AUDUSD M1

[LondonScalper]

Gap fill on AUDUSD M1 is on my research list, not my live playbook. Notes so far below.

Definition I use: the Monday open is at least 15 pips away from Friday's close, and London opens with the gap still unfilled. The question is whether price fills it during London, and whether a scalp toward the fill is worth the cost.

From roughly forty cases over the last year in my notes: the gap was filled during London a bit over half the time. That sounds tradeable, but the fills that did happen often came late, after a move in the opposite direction first. A tight scalp stop would have been taken out in many of them before the fill.

The other problem is spread. Monday morning AUDUSD spreads on my account are noticeably wider in the first hour, and most gap discussions online ignore that completely. Once I subtract cost, the edge in my sample gets very small.

So for now it stays a tag in the journal. If I see a gap, I note it and record what happened, but I do not trade it. Maybe another year of data changes that. Does anyone trade AUD gaps with real money and good results?

[Fairman]

Definition I use: the Monday open is at least 15 pips away from Friday's close, and London opens with the gap still unfilled.

One thing that can distort the sample is which price counts as the Monday open. The first quotes after the Sunday open, around 22:00 London, often come from only a few liquidity providers with very wide spreads, and on some brokers the first tick sits noticeably away from where the market settles a few minutes later. So a gap measured from the first M1 bar can look bigger than the gap that existed in tradeable prices. Measuring it from the mid price after the first 15 or 30 minutes, once spreads are near normal, tends to shrink a few gaps below the 15-pip threshold and might change the fill rate. The other split worth doing is by cause: a gap from weekend news on China or Australia is a different animal from one that is just thin-market drift.

[Shadow Trader]

LondonScalper, your fill rate a bit over half matches the 43 Monday gaps I tagged last year, but once I applied Fairman's mid price filter the tradeable sample shrank a lot. Measuring from the first M1 tick gave me 43 gaps of 15 pips or more, while the mid after the first 30 minutes left only 29.

Of those 29, London filled 16, so still roughly half, and 9 of the 16 filled only after an adverse swing of at least 8 pips. A tight scalp stop would have been gone before the fill on those.

img_13420_r1.png (51.75 KiB) Viewed 6 times

The illustration shows first tick versus settled mid counts, and how many of the settled gaps actually filled in London. Spread in the first hour remains the quiet killer in both versions.

I also split by cause as Fairman suggested. China weekend headlines filled less often than thin market drift gaps in my notes, though the China sample is only eleven cases.

Are you still leaving these as journal tags only, or has another year of data pulled any of them onto a watchlist with wider stops? Mine stay tags for now.