Order Book Imbalance and Nonlinear Weighted Mid-Price Estimation

Summary

This high-frequency market microstructure analysis uses Binance best-bid and best-ask price and size data for YGG on a volatile day. It examines how quote spread and top-of-book quantities changed, then defines imbalance from the relative sizes of the best bid and ask. The observed next-interval mid-price changes generally rise with bid-side imbalance, supporting the idea that a thinner side of the book may be more easily consumed by incoming orders.

The article compares the ordinary midpoint with a weighted midpoint that shifts according to the spread and imbalance. It finds that the simple linear adjustment reduces the relationship between imbalance and subsequent price change, but residual patterns remain. It proposes a nonlinear adjustment, including an odd-power form, and reports that this fits the sample somewhat better. The analysis is exploratory: it uses only level-one quotes from one asset-day, does not establish live profitability, and says that deeper book data and actual trades should be used to evaluate a practical fair-price estimate.

Key ideas

• Best-quote size imbalance contains information about the direction of the next short-interval mid-price move in the sample.
• A weighted midpoint adjusts the ordinary midpoint using the bid-ask spread and relative queue sizes.
• The simple linear adjustment leaves systematic residual patterns, motivating a nonlinear imbalance correction.
• The proposed nonlinear form is a sample fit and is not presented as ready for live trading.
• The study is limited to level-one quotes for one asset-day and calls for validation using deeper books and trades.

Tags

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