Do VLMs Truly "Read" Candlesticks? A Multi-Scale Benchmark for Visual Stock Price Forecasting

• URL: http://arxiv.org/abs/2604.12659v1
• Date: Tue, 14 Apr 2026 12:26:34 GMT
• Title: Do VLMs Truly "Read" Candlesticks? A Multi-Scale Benchmark for Visual Stock Price Forecasting
• Abstract summary: Vision-language models (VLMs) are increasingly applied to visual stock price forecasting, yet existing benchmarks inadequately evaluate their understanding of stock price in candlestick charts.<n>We construct a multi-scale candlestick charts dataset and a standardized evaluation framework to assess VLMs' ability to utilize multi-scale visual market signals.
• Score: 2.2212162466658882
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-language models(VLMs) are increasingly applied to visual stock price forecasting, yet existing benchmarks inadequately evaluate their understanding of stock price in candlestick charts. First, prior studies fail to isolate VLMs' comprehension of visual inputs genuinely improves predictive performance and whether VLMs truly comprehend candlestick patterns. Further, most existing datasets and evaluation setups are designed around single-period or tabular inputs. However, human analysts strongly rely on multi-scale candlestick charts, where longer-term horizons capture trend direction and shorter-term horizons provide cues for inflection points, making it difficult to systematically assess VLMs' ability to integrate short-term and long-term visual market dynamics. To bridge this gap, we construct a multi-scale candlestick charts dataset and a standardized evaluation framework to assess VLMs' ability to utilize multi-scale visual market signals. Evaluation combines confusion-matrix-based diagnostics with information coefficient(IC) time series metrics and includes XGBoost as a feature-based temporal baseline. Using this dataset, we benchmark representative VLMs and analyze their ability to leverage multi-scale stock price data. Experimental results show that most VLMs perform well only under persistent uptrend or downtrend conditions, while exhibiting weak predictive capability in more common market scenarios. We also identify significant prediction biases and limited sensitivity to explicitly specified forecast horizons in prompts, indicating inherent limitations in precise temporal reasoning.

Related papers

• TimeVista: Exploring and Exploiting Vision-Language Models as Judges for Time Series Forecasting [51.59529402134439]
  ''LLM-as-a-Judge'' paradigm has revolutionized text evaluation by providing flexible, human-aligned judgment.<n>We propose a novel framework integrating micro- and macro-level judgments informed by contextual information.<n>Extensive meta-evaluations demonstrate that Vision-Language Models (VLMs) are highly reliable judges.
  arXiv  Detail & Related papers  (2026-06-15T03:39:55Z)
• Vision-aligned Latent Reasoning for Multi-modal Large Language Model [82.26044667101011]
  Vision-aligned Latent Reasoning (VaLR) is a framework that dynamically generates vision-aligned latent tokens before each Chain of Thought reasoning step.<n>VaLR is trained to preserve visual knowledge during reasoning by aligning intermediate embeddings of MLLM with those from vision encoders.
  arXiv  Detail & Related papers  (2026-02-04T12:04:02Z)
• Look-Ahead-Bench: a Standardized Benchmark of Look-ahead Bias in Point-in-Time LLMs for Finance [0.0]
  We introduce Look-Ahead-Bench, a standardized benchmark measuring look-ahead bias in Point-in-Time (PiT) Large Language Models (LLMs)<n>We analyze performance decay across temporally distinct market regimes, incorporating several quantitative baselines to establish performance thresholds.<n>Results reveal significant lookahead bias in standard LLMs, as measured with alpha decay, unlike Pitinf models.
  arXiv  Detail & Related papers  (2026-01-20T09:23:51Z)
• When to Think and When to Look: Uncertainty-Guided Lookback [41.85837765999506]
  We show that more thinking is not always better; long chains often yield long wrong trajectories that ignore the image.<n>A deeper analysis reveals that certain short lookback phrases, which explicitly refer back to the image, are strongly enriched in successful trajectories.<n>We propose uncertainty guided lookback, a training free decoding strategy that combines an uncertainty signal with adaptive lookback prompts and breadth search.
  arXiv  Detail & Related papers  (2025-11-19T17:01:02Z)
• Why Bonds Fail Differently? Explainable Multimodal Learning for Multi-Class Default Prediction [4.838838129678638]
  We propose a novel framework for multi-class bond default prediction.<n>LOT integrates numerical time-series (financial/macroeconomic indicators) and unstructured data (bondes)<n>It uses Time-Aware LSTM to handle irregular sequences, and adopts soft clustering and multi-level attention to boost interpretability.
  arXiv  Detail & Related papers  (2025-09-13T03:42:34Z)
• VISTA: Vision-Language Inference for Training-Free Stock Time-Series Analysis [0.0]
  We introduce VISTA (Vision-Language Inference for Stock Time-series Analysis), a training-free framework for multi-modal stock forecasting.<n>We benchmark VISTA against standard baselines, including ARIMA and text-only LLM-based prompting methods.<n>We show that VISTA outperforms these baselines by up to 89.83%, demonstrating the effectiveness of multi-modal inference for stock time-series analysis.
  arXiv  Detail & Related papers  (2025-05-24T07:20:14Z)
• Protecting multimodal large language models against misleading visualizations [94.71976205962527]
  We show that questionanswering (QA) accuracy on misleading visualizations drops on average to the level of the random baseline.<n>We introduce the first inference-time methods to improve QA performance on misleading visualizations, without compromising accuracy on non-misleading ones.<n>We find that two methods, table-based QA and redrawing the visualization, are effective, with improvements of up to 19.6 percentage points.
  arXiv  Detail & Related papers  (2025-02-27T20:22:34Z)
• Agent Trading Arena: A Study on Numerical Understanding in LLM-Based Agents [69.58565132975504]
  Large language models (LLMs) have demonstrated remarkable capabilities in natural language tasks.<n>We present the Agent Trading Arena, a virtual zero-sum stock market in which LLM-based agents engage in competitive multi-agent trading.
  arXiv  Detail & Related papers  (2025-02-25T08:41:01Z)
• Elevating Visual Perception in Multimodal LLMs with Visual Embedding Distillation [109.5893580175657]
  In recent times, the standard practice for developing MLLMs is to feed features from vision encoder(s) into the LLM and train with natural language supervision.<n>This approach often causes models to lean towards language comprehension and undermine the rich visual perception signals present in the data.<n>We propose VisPer-LM, the first approach that infuses visual perception knowledge from expert vision encoders into the LLM's hidden representations.
  arXiv  Detail & Related papers  (2024-12-12T18:55:18Z)
• Diffusion Variational Autoencoder for Tackling Stochasticity in Multi-Step Regression Stock Price Prediction [54.21695754082441]
  Multi-step stock price prediction over a long-term horizon is crucial for forecasting its volatility. Current solutions to multi-step stock price prediction are mostly designed for single-step, classification-based predictions. We combine a deep hierarchical variational-autoencoder (VAE) and diffusion probabilistic techniques to do seq2seq stock prediction. Our model is shown to outperform state-of-the-art solutions in terms of its prediction accuracy and variance.
  arXiv  Detail & Related papers  (2023-08-18T16:21:15Z)
• Long Short-Term Memory Neural Network for Financial Time Series [0.0]
  We present an ensemble of independent and parallel long short-term memory neural networks for the prediction of stock price movement. With a straightforward trading strategy, comparisons with a randomly chosen portfolio and a portfolio containing all the stocks in the index show that the portfolio resulting from the LSTM ensemble provides better average daily returns and higher cumulative returns over time.
  arXiv  Detail & Related papers  (2022-01-20T15:17:26Z)
• Trade When Opportunity Comes: Price Movement Forecasting via Locality-Aware Attention and Iterative Refinement Labeling [11.430440350359993]
  We propose LARA, a novel price movement forecasting framework with two main components. LA-Attention extracts potentially profitable samples through masked attention scheme. RA-Labeling refines the noisy labels of potentially profitable samples. LARA significantly outperforms several machine learning based methods on the Qlib quantitative investment platform.
  arXiv  Detail & Related papers  (2021-07-26T05:52:42Z)
• Deep Stock Predictions [58.720142291102135]
  We consider the design of a trading strategy that performs portfolio optimization using Long Short Term Memory (LSTM) neural networks. We then customize the loss function used to train the LSTM to increase the profit earned. We find the LSTM model with the customized loss function to have an improved performance in the training bot over a regressive baseline such as ARIMA.
  arXiv  Detail & Related papers  (2020-06-08T23:37:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.