forex-scalping.com

From my own raw-account log, not a broker brochure: EURUSD all-in cost (spread + commission + median slip on market orders) was lowest for me between roughly 07:30 and 10:30 London on normal weeks, with a second decent pocket after US open once spreads settled.

Worst pockets: the minute around Tier-1 releases, the rollover window, and very quiet mid-August mornings where spread looked fine but slip on stops was ugly.

I do not treat this as eternal truth. I re-check a month of deals quarterly. The point of logging hours is to stop trading my "favourite" setup at my most expensive clock time just because I am awake. Cost is part of the session filter, same as Asia range width.

When a month shows a new expensive hour - for example a lunch liquidity hole - that hour gets demoted on the roster until the log improves. The roster follows the cost data. I also exclude holidays and DST-shift weeks from the baseline sample so the "best hours" claim stays honest.

My log agrees on the morning pocket, with one difference. The first 10 to 15 minutes after 08:00 London were noticeably worse for me than 08:15 to 10:30, mostly slip on market entries rather than the quoted spread.

For scale, my all-in on EURUSD raw sits around 0.8 to 0.9 pips in that window, and commission is about 0.7 of that. Spread and slip are the part that moves by hour, so that's what my hourly table actually tracks.

The rollover window is the ugly one. From roughly 21:55 to 22:15 London I've seen quoted spreads of 1.5 pips and more, so anything still open with a tight stop gets a free lesson.

One thing I'd add to your method: log missed limit orders by hour too. Limits don't slip, but in some hours a fair share never fill, and that cost doesn't show up on the deals tab at all.

Fully agree on dropping DST-shift weeks from the baseline. The US data time moves an hour in London terms, so those weeks blur the "after US open" pocket.

What sample size do you use per hour before you trust a bucket? I go for at least 40 fills.

PropScalpDesk, I split my hourly table into the fixed part and the moving part, and it changed how I read the cost ranking. Commission on my raw account is a flat 0.70 pips round turn, so every hour starts from the same floor.

What moves is spread plus slip on top of it. Over my last three months that variable part ran from 0.12 pips in the 09:00 London hour to 0.40 in the 13:00 hour, where the 13:30 US data sits. The chart below is an illustration of that split by hour.
So the gap between my best and worst hour is more than three times on the part I can actually control. Seen as totals of 0.82 against 1.10 pips it looks small, which is why blended averages hide it.

The afternoon pocket you mention shows up too, at 0.86 in the 15:00 hour once the US open settles. The 12:00 hour at 0.98 is the lunch hole that keeps getting demoted on my roster.

Do you rank hours on the total or on the variable part only? On a fixed commission account I think the variable part is the honest comparison.