Order Book Imbalance and Nonlinear Microprice Estimation

Summary

This article examines whether top-of-book bid and ask quantities can improve estimates of short-term price movement. Using Binance YGG book ticker data from a volatile day, it defines quote imbalance from the relative sizes of the best bid and ask queues, then compares that signal with subsequent mid-price changes. The observed relationship is directional: as imbalance shifts toward one side, the next mid-price change tends to favor that side.

The article then adjusts the ordinary midpoint by the spread and imbalance to form a weighted mid-price. It finds that this reduces the apparent dependence of future price changes on imbalance, though residual patterns remain, especially near the extremes. A nonlinear correction is proposed, with an eighth-power term fitting the example better than the simple linear adjustment. The author presents this as an exploratory approach rather than a live-ready model. The analysis uses only the best bid and ask, and the article says practical estimates should also be checked against trade data and deeper available book information; it points to Markov-chain microprice work as a further direction.

Key ideas

• Top-of-book queue sizes can provide information about the direction of near-term price changes.
• The ordinary midpoint ignores queue imbalance and may be a weaker short-horizon price estimate.
• A spread-scaled imbalance adjustment creates a weighted midpoint, but a linear adjustment leaves residual structure.
• The example proposes a nonlinear correction, while cautioning that it is not yet suitable for live trading.
• The analysis uses only top-level quotes and should be validated against trades and fuller market data.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.