The problem with "average" in average daily range

[Shadow Trader]

Average daily range figures, like 80 pips on EURUSD, are published widely. They guide targets and stops. But averages mislead.

The distribution of daily ranges is skewed. Many days fall around 50 to 70 pips, a few reach 150 or more, and those outliers lift the average.

A median gives a better sense of a typical day. If the average is 80 and the median is 65, then most days fall short of the average.

Targets set at the average will rarely be reached. Daily range also depends on weekday, session and event calendar. A day with major data is a different animal from a quiet Tuesday.

For scalpers, intraday range by hour is more useful than daily figures. It shows how much movement to expect in the exact window being traded.

Computing medians and percentiles from the last 100 sessions takes minutes in a spreadsheet.

How are range expectations set here, from averages or from the actual distribution?

[Amanhasnoname]

I worked this out for EURUSD a while ago and the gap was bigger than I expected. My hundred session average was around 78 pips, but the median was closer to 63, and fewer than four days in ten actually reached the average. I had been setting my daily expectations off a number that was being pulled up by a handful of data days.

The more useful change was doing it by hour. For my window, 08:00 to 10:00 London, the median range was only about 22 pips, and the top ten percent of sessions were more than twice that. Knowing the median stopped me holding for targets that the window rarely delivers, while the top end told me when a morning was unusual and worth staying with.

I also split it by weekday and by whether there was high-impact data. Mondays and quiet Fridays had noticeably smaller medians. It took one spreadsheet with a few pivot tables, and it changed my targets more than any indicator I've ever tried.