CTBench: Cryptocurrency Time Series Generation Benchmark
Not on the map. Why?
Abstract (arXiv)
Synthetic time series are essential tools for data augmentation, stress testing, and algorithmic prototyping in quantitative finance. However, in cryptocurrency markets, characterized by 24/7 trading, extreme volatility, and rapid regime shifts, existing Time Series Generation (TSG) methods and benchmarks often fall short, jeopardizing practical utility. Most prior work (1) targets non-financial or traditional financial domains, (2) focuses narrowly on classification and forecasting while neglecting crypto-specific complexities, and (3) lacks critical financial evaluations, particularly for trading applications. To address these gaps, we introduce textsfCTBench, the first comprehensive TSG benchmark tailored for the cryptocurrency domain. textsfCTBench curates an open-source dataset from 452 tokens and evaluates TSG models across 13 metrics spanning 5 key dimensions: forecasting accuracy, rank fidelity, trading performance, risk assessment, and computational efficiency. A key innovation is a dual-task evaluation framework: (1) the Predictive Utility task measures how well synthetic data preserves temporal and cross-sectional patterns for forecasting, while (2) the Statistical Arbitrage task assesses whether reconstructed series support mean-reverting signals for trading. We benchmark eight representative models from five methodological families over four distinct market regimes, uncovering trade-offs between statistical fidelity and real-world profitability. Notably, textsfCTBench offers model ranking analysis and actionable guidance for selecting and deploying TSG models in crypto analytics and strategy development.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 44 of its 83 references that are in the map. First posted on arXiv 3 Aug 2025.
Builds on
44 of 83 references.
References · 44
- 2017Attention Is All You Needcited by 60,361
- 2022Generative Adversarial Networkscited by 5,627
- 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
- 2018Neural Ordinary Differential Equationscited by 3,779
- 2017Real-valued (Medical) Time Series Generation with Recurrent Conditional GANscited by 333
- 2024TFB: Towards Comprehensive and Fair Benchmarking of Time Series Forecasting Methodscited by 249
- 2018Adversarial Audio Synthesiscited by 225
- 2024Diffusion-TS: Interpretable Diffusion for General Time Series Generationcited by 184
- 2021Neural SDEs as Infinite-Dimensional GANscited by 161
- 2021TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generationcited by 142
- 2016C-RNN-GAN: Continuous recurrent neural networks with adversarial trainingcited by 132
- 2019Quant GANs: Deep Generation of Financial Time Seriescited by 108
32 more
- 2020COT-GAN: Generating Sequential Data via Causal Optimal Transportcited by 71
- 2021Sig-Wasserstein GANs for Time Series Generationcited by 60
- 2021PSA-GAN: Progressive Self Attention GANs for Synthetic Time Seriescited by 58
- 2023On the Constrained Time-Series Generation Problemcited by 54
- 2020Time Series Domain Adaptation via Sparse Associative Structure Alignmentcited by 51
- 2023Vector Quantized Time Series Generation with a Bidirectional Prior Modelcited by 50
- 2023Generative Modeling of Regular and Irregular Time Series Data via Koopman VAEscited by 49
- 2023Causal Recurrent Variational Autoencoder for Medical Time Series Generationcited by 45
- 2022GT-GAN: General Purpose Time Series Synthesis with Generative Adversarial Networkscited by 44
- 2024Utilizing Image Transforms and Diffusion Models for Generative Modeling of Short and Long Time Seriescited by 43
- 2022TTS-GAN: A Transformer-based Time-Series Generative Adversarial Networkcited by 41
- 2022Deep Latent State Space Models for Time-Series Generationcited by 38
- 2020Modeling Continuous Stochastic Processes with Dynamic Normalizing Flowscited by 37
- 2020Conditional GAN for timeseries generationcited by 37
- 2018T-CGAN: Conditional Generative Adversarial Network for Data Augmentation in Noisy Time Series with Irregular Samplingcited by 29
- 2021Towards Generating Real-World Time Series Datacited by 26
- 2023TSGBench: Time Series Generation Benchmarkcited by 26
- 2021Reinforcement Learning for Quantitative Tradingcited by 24
- 2023Reliable Generation of Privacy-preserving Synthetic Electronic Health Record Time Series via Diffusion Modelscited by 24
- 2025FinTSB: A Comprehensive and Practical Benchmark for Financial Time Series Forecastingcited by 24
- 2022Generating multivariate time series with COmmon Source CoordInated GAN (COSCI-GAN)cited by 23
- 2023Regular Time-series Generation using SGMcited by 23
- 2023TSGM: A Flexible Framework for Generative Modeling of Synthetic Time Seriescited by 18
- 2025Population Aware Diffusion for Time Series Generationcited by 16
- 2020DoubleEnsemble: A New Ensemble Method Based on Sample Reweighting and Feature Selection for Financial Data Analysiscited by 12
- 2022Time-series Transformer Generative Adversarial Networkscited by 10
- 2024TimeLDM: Latent Diffusion Model for Unconditional Time Series Generationcited by 10
- 2023Time-series Generation by Contrastive Imitationcited by 9
- 2023Towards Generalizable Reinforcement Learning for Trade Executioncited by 5
- 2021Conditional Loss and Deep Euler Scheme for Time Series Generationcited by 4
- 2024Towards Controllable Time Series Generationcited by 4
- 2025FinTSBridge: A New Evaluation Suite for Real-world Financial Prediction with Advanced Time Series Modelscited by 1
Built on by
Citing papers per year: 2025: 2; 2026: 1.
3 papers.