IC Markets cTrader DOM screenshots worth studying

[LondonScalper]

cTrader DOM screenshots are study material, not a crystal ball.

I capture the DOM on EURUSD in the first ten minutes after 08:00 and again near 09:30 on quiet days, then compare how quickly top-of-book size refreshes. I am looking for when the book is too fast to read versus when thinning on one side comes before a small lift. Sample size is still too small for rules, so screenshots stay in a research folder, not in the live playbook.

What I have noticed so far, with heavy caveats: in the first minute after 08:00 the book changes faster than I can read it, so any idea of reacting to it by hand is fantasy. Around 09:30 on a quiet day it slows down enough that you can see size sitting at a level for several seconds. Whether that size means anything is a different question. Retail DOM shows one liquidity provider's view, not the whole market.

Until I have a few hundred tagged observations, the DOM does not change a single entry criterion. Curiosity is allowed. Premature certainty is not.

Anyone else keeping DOM screenshots? What do you tag them with?

[Shadow Trader]

A still screenshot can't show the thing you're measuring, which is refresh speed. I ran into that within a week and switched to logging the book instead, using a small cTrader Automate bot that writes every depth update to a CSV with a millisecond timestamp.

That turned the book slowing down near 09:30 into a number. On my feed, top of book on EURUSD changed roughly 40 times a second in the first minute after 08:00 and closer to 5 a second on quiet mid-morning stretches, though that's one broker and a few weeks of data.

For tags I keep it boring: time, spread, the top three levels of size on each side, and where mid was 30 and 120 seconds later. I also flag anything within 15 minutes of a scheduled release, because those rows behave differently and would swamp the rest.

One edge case worth tagging separately is size that gets hit and reappears at the same price straight away. On a retail feed that's usually the provider refreshing its quote, not someone defending a level, and counting it as a wall skews the sample.

How long does size need to sit before you call it sitting? I settled on 3 seconds, mostly because anything shorter was impossible to tell apart from noise.

[Shadow Trader]

LondonScalper, I charted the morning refresh curve from the same CSV log I mentioned before. On quiet days top of book on EURUSD starts near 42 updates a second at 08:00, falls to about 18 by 08:30, and settles near 6 by 09:30, which matches the feel of your screenshots with a number under it.

The chart below is an illustration of that median curve across twelve quiet mornings. The steep drop in the first half hour is why I treat anything before 08:20 as research only, never as a live read.

img_13300_r2.png (76.25 KiB) Viewed 4 times

A tag I would add to your list is cancel to replace ratio at the best bid or ask. When that ratio spikes without a release nearby, my mid often drifted 2 to 3 pips within two minutes on the sample I have.

It is still curiosity, not an entry rule. Do you already mark cancel heavy minutes, or would that clutter the folder before the sample is large enough?