Adapting the GLFT Market-Making Model for Grid Quotes

Summary

This tutorial applies the Guéant–Lehalle–Fernandez-Tapia market-making model to grid quoting. It derives bid and ask quote depths from a fair price, volatility, trading intensity, and inventory. The resulting quotes combine a half-spread with an inventory-sensitive skew: inventory shifts the bid and ask in opposite directions. The article then estimates the exponential trading-intensity curve from observed market-trade arrival depths, fitting its parameters with log-linear regression, and estimates volatility from changes in mid-price.

A replay example uses crypto market data and a 100-millisecond decision interval, then discusses plugging the estimates into a grid strategy. The evidence is illustrative: the fitted intensity curve is reported to overstate activity near the mid-price and understate it farther away, leading to a refit over shallow quote depths. Queue position is omitted, and the text notes a half-tick versus tick-unit inconsistency in one measurement routine. It also cautions that historical intensity and volatility, mid-price fair value, and inventory-only risk are limited assumptions; improved quotes may require forecasts, richer risk controls, correlated-asset pricing, and hedging.

Key ideas

• The GLFT approximation sets quote depth using trading intensity, volatility, and the market maker’s inventory.
• Inventory creates quote skew by moving bid and ask depths in opposite directions.
• The tutorial estimates exponential trading intensity from trade arrival depths and calibrates it with log-linear regression.
• Its example fits the full depth range poorly and refits only the shallow range where quotes are more likely to sit.
• The simplified replay omits queue position and relies on historical conditions and mid-price fair value.

Tags

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