DiffLOB: Diffusion Models for Counterfactual Generation in Limit Order Books

Zhuohan Wang, Carmine Ventre · 2026

Read DiffLOB: Diffusion Models for Counterfactual Generation in Limit Order Books by Zhuohan Wang, Carmine Ventre (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-03776.txt.

Description

A limit order book records the standing bids and offers behind every trade, and simulating what might have happened would help stress-test trading plans. Zhuohan Wang and Carmine Ventre propose DiffLOB in a 2026 paper, a diffusion model that generates order book trajectories conditioned on future market regimes. Existing generators either replay typical dynamics or need a live interacting agent.

DiffLOB conditions on trend, volatility, liquidity, and order-flow imbalance, so it can answer what-if questions: how would the book evolve under a different regime? The authors build a three-part evaluation for such counterfactual generation. It checks controllable realism, whether intervening on a regime changes the dynamics consistently, and whether the synthetic trajectories improve downstream regime prediction.

Catalog record

CREATORZhuohan Wang, Carmine Ventre
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-03776

Browse the full library · About the Tower of Babel