Tick volume divergences are mostly noise, here is why
Posted: Sat Oct 31, 2026 6:33 am
Plenty of strategies treat a price high on falling tick volume as a warning sign. The logic sounds neat: fewer participants pushing the move, so it must be weak.
The trouble is how many things lower tick counts without any change in participation. A quieter feed stretch, a platform lag, a broker pausing updates during a server hiccup, or simply the natural drop in activity as a session winds down all produce the same picture.
Testing it properly means separating those effects. Compare divergence signals with the same time of day, controlling for session, and the apparent edge often shrinks to almost nothing. The pattern looks reliable in hindsight because every reversal gets labelled a divergence afterward, while the many divergences that led nowhere get forgotten.
A fairer use of tick volume is as a filter for conditions, such as skipping setups when activity is unusually low, rather than as a predictor of direction.
Has a tick volume divergence rule survived a proper test across several months of data?
The trouble is how many things lower tick counts without any change in participation. A quieter feed stretch, a platform lag, a broker pausing updates during a server hiccup, or simply the natural drop in activity as a session winds down all produce the same picture.
Testing it properly means separating those effects. Compare divergence signals with the same time of day, controlling for session, and the apparent edge often shrinks to almost nothing. The pattern looks reliable in hindsight because every reversal gets labelled a divergence afterward, while the many divergences that led nowhere get forgotten.
A fairer use of tick volume is as a filter for conditions, such as skipping setups when activity is unusually low, rather than as a predictor of direction.
Has a tick volume divergence rule survived a proper test across several months of data?