---
title: Transformers versus LSTMs for electronic trading
url: https://www.ml-quant.com/papers/arxiv/2309.11400/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-09-26
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: arXiv:2309.11400
source_url: https://arxiv.org/abs/2309.11400
featured: 2023-09-21
citations: 33
topic: Trading, Microstructure & Execution
---


# Transformers versus LSTMs for electronic trading

A comparison study of LSTM-based and Transformer-based models for financial time series prediction found that LSTM-based models perform better in difference sequence prediction, despite Transformer-based models having limited advantages in absolute price sequence prediction.

- Source: https://arxiv.org/abs/2309.11400
- Identifier: arXiv:2309.11400
- Released: 2023-09-20
- First featured: Quant Letter No. 16 (2023-09-21): https://www.ml-quant.com/issues/2023-09-21/
- Citations (Semantic Scholar): 33
- Published in: not yet
- Topic: Trading, Microstructure & Execution

## Related

- [Deep attentive survival analysis in limit order books: estimating fill probabilities with convolutional-transformers](https://www.ml-quant.com/papers/arxiv/2306.05479/): A deep learning method outperforms other approaches in estimating filltimes of limit orders.
- [VWAP Execution with Signature-Enhanced Transformers: A Multi-Asset Learning Approach](https://www.ml-quant.com/papers/arxiv/2503.02680/): A new method for Volume Weighted Average Price (VWAP) execution has been suggested, using a single neural network across multiple assets, providing a more scalable and effective solution than traditional asset-specific models.
- [Detecting Multilevel Manipulation from Limit Order Book via Cascaded Contrastive Representation Learning](https://www.ml-quant.com/papers/arxiv/2508.17086/): The study suggests a learning framework to enhance the detection of trade-based manipulation in financial markets, with Transformer-based architectures proving most successful.
- [Transformers vs LSTMs in Electronic Trading](https://www.ml-quant.com/papers/ssrn/4577922/): The research compares LSTM and Transformer models in financial prediction tasks, finding out if Transformer models can surpass LSTM in financial time series prediction.
- [Deep Unsupervised Anomaly Detection in High-Frequency Markets](https://www.ml-quant.com/papers/ssrn/4502662/): A new anomaly detection framework for stock trading data uses a modified Transformer autoencoder to spot fraudulent time series.
- [FAST: Efficient Action Tokenization for Vision-Language-Action Models](https://www.ml-quant.com/papers/arxiv/2501.09747/): A new tokenization scheme, Frequency-space Action Sequence Tokenization (FAST), has been proposed for robot actions, facilitating the training of vision-language action policies for complex and high-frequency tasks.