Tick imbalance: counting upticks against downticks

[Shadow Trader]

A simple way to measure short-term pressure is to count how many ticks moved the price up compared with how many moved it down over the last N seconds. If upticks dominate, buyers are pressing. If downticks dominate, sellers are.

It sounds crude, and it is, but it captures something real: the direction of quote changes reflects the direction in which liquidity providers are repricing. A string of ask increases with the bid unchanged says something different from a symmetrical flow.

Problems appear quickly. The count depends on the data feed and the broker's tick generation. Quote flicker inflates both sides. And the signal decays fast, with predictive value measured in seconds.

A careful version uses only ticks where the mid price actually changed, ignores updates that only widened the spread, and normalizes by the total count.

Even then, expecting a clean edge is optimistic. Most of what shows up is already reflected in price by the time an order can be placed. The measure serves better as a filter, avoiding entries when imbalance clearly points the opposite way.

Is a tick imbalance filter worth the bother, or does it just add noise?