Volume bars: when the clock is the wrong measure
Instead of cutting bars by time, a volume bar closes after a fixed amount of trading. In forex, where true volume is hidden, tick counts stand in.
The appeal is statistical. Time bars contain a variable amount of activity, with some holding a flurry of ticks and others almost none. Volume bars contain a roughly constant amount, so each bar carries similar information, and the returns tend to behave better for analysis.
For scalping, they smooth out the dead periods and zoom in on the active ones. A signal based on volume bars responds to market activity rather than the clock.
The weak points are the usual ones. Tick counts depend on the broker's feed. Platforms need custom tools to build them. And live behaviour can differ from historical construction if the feed changes.
Anyone curious can test a basic idea on both bar types and compare the stability of the results.
A neat property is that volume bars normalise the data. During a busy minute, many bars form. During a quiet one, few. A fixed-size move in the bar series then represents a roughly constant amount of activity.
Have volume or tick bars given a cleaner view than time bars?
Volume bars: when the clock is the wrong measure
Volume bars: when the clock is the wrong measure
It’s Fairman 
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Volume bars: when the clock is the wrong measure
I spent a few weeks building tick bars for GBPUSD because the dead stretch between the New York close and Tokyo kept generating junk signals on my one-minute chart. I settled on bars of two hundred ticks after trying smaller sizes, which gave roughly the same number of bars as the one-minute chart across a full day, but with very few during the quiet hours and many during London.
Running the same moving-average pullback rules on both, the tick-bar version took about a third fewer trades and had a smaller worst losing streak. The average result per trade was only slightly better, so most of the gain came from simply not trading the empty periods, which a time filter could have done as well.
The catch was the feed. When I moved the test to a second broker, the tick counts per day were quite different and the bars no longer lined up, so the settings had to be redone. I still use them for research, but I trade from time bars with a session filter because it is easier to keep consistent.
Running the same moving-average pullback rules on both, the tick-bar version took about a third fewer trades and had a smaller worst losing streak. The average result per trade was only slightly better, so most of the gain came from simply not trading the empty periods, which a time filter could have done as well.
The catch was the feed. When I moved the test to a second broker, the tick counts per day were quite different and the bars no longer lined up, so the settings had to be redone. I still use them for research, but I trade from time bars with a session filter because it is easier to keep consistent.