Deep models predict market impact of trading scenarios without retraining
Repurposing Deep Limit Order Book Forecasting for Scenario-Conditioned Market Impact Modeling
Machine LearningArtificial Intelligence
Summary
Predicting how trades affect stock prices is tricky because markets move in complex ways. The authors show that models trained to forecast short-term market data can also estimate the impact of hypothetical trading actions. Their approach injects realistic fake trading messages into the data and compares predictions before and after. This lets them measure how certain trades might change the market quickly and accurately, all without having to retrain the model.
What this means in practice
- •For quantitative traders: Estimate short-term price effects of potential large trades using pretrained order book models without extra training.
- •For financial risk teams: Assess how hypothetical order book events might change market prices to improve scenario analysis and risk management.
Authors
Eljas Linna, Kestutis Baltakys, Derrick Manoharan, Alexandros Iosifidis, Juho Kanniainen
Abstract
Deep Limit Order Book forecasting models capture nonlinear market dynamics, but their ability to quantify the effects of counterfactual order book messages has not been systematically validated. We introduce a model-agnostic framework that compares a trained forecaster's predictive distributions before and after injecting mechanically valid counterfactual messages, defining short-horizon model-implied market impact. A Transformer-based forecaster recovered scenario rankings with a Spearman correlation of 0.99 and 97.2% directional agreement with realized historical outcomes among non-neutral scenarios. Observation-level analysis further showed that estimated impacts captured incremental sequence-dependent variation beyond scenario identity and the pre-event forecast. These results provide evidence that pretrained Limit Order Book forecasters can be repurposed for scenario-conditioned response modeling without retraining.