Paper: arXiv 2511.15960

Authors: Gabriel M. Arantes, Richard F. Pinto, Bruno L. Dalmazo, Eduardo N. Borges, Giancarlo Lucca, Viviane L. D. de Mattos, Fabian C. Cardoso, Rafael A. Berri

Abstract

Binary options trading is often marketed as a field where predictive models can generate consistent profits. However, the inherent randomness and stochastic nature of binary options make price movements highly unpredictable, posing significant challenges for any forecasting approach. This study demonstrates that machine learning algorithms struggle to outperform a simple baseline in predicting binary options movements. Using a dataset of EUR/USD currency pairs from 2021 to 2023, we tested multiple models, including Random Forest, Logistic Regression, Gradient Boosting, and k-Nearest Neighbors (kNN), both before and after hyperparameter optimization. Furthermore, several neural network architectures, including Multi-Layer Perceptrons (MLP) and a Long Short-Term Memory (LSTM) network, were evaluated under different training conditions. Despite these exhaustive efforts, none of the models surpassed the ZeroR baseline accuracy, highlighting the inherent randomness of binary options. These findings reinforce the notion that binary options lack predictable patterns, making them unsuitable for machine learning-based forecasting.

Complexity vs Empirical Score

  • Math Complexity: 2.5/10
  • Empirical Rigor: 7.5/10
  • Quadrant: Street Traders — empirical and applied, lower math complexity

Why this score: The paper demonstrates a rigorous empirical approach with a substantial real-world dataset, extensive model testing, and optimization, though it lacks the advanced mathematical derivations typical of theoretical finance.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Can ML predict<br>binary options movements?"] --> B["Data<br>EUR/USD 2021-2023"]
  B --> C["Methodology<br>Train/Test Models"]
  C --> D["Models Tested<br>Random Forest, LR, Gradient Boosting,<br>kNN, MLP, LSTM"]
  D --> E["Baseline Comparison<br>ZeroR Accuracy"]
  E --> F["Key Outcome<br>Models failed to beat<br>baseline; high randomness"]
  F --> G["Conclusion<br>Binary options are<br>inherently unpredictable"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper evaluates a range of machine learning models against a ZeroR baseline to assess predictability of binary options movements.

Source sentence
This study demonstrates that machine learning algorithms struggle to outperform a simple baseline in predicting binary options movements.

Assumes 4

  • Binary options price movements are entirely random, providing no predictable patterns. market
    Source sentence
    the hypothesis that binary options are entirely random
  • Technical indicators (SMA and RSI) are sufficient features to capture any potential predictive signal. data
    Source sentence
    We derived a set of technical indicators: Simple Moving Averages (SMA) with windows of 2, 3, 4, and 5 minutes
  • Training and test sets are split chronologically to avoid information leakage. data
    Source sentence
    The dataset was divided into a training set (2021–2022, N = 742,240) and a testing set (2023, N = 322,572) to ensure models learn from historical patterns without information leakage.
  • Accuracy is an appropriate metric for evaluating binary classification performance. statistical
    Source sentence
    Accuracy was the main performance metric, enabling direct comparison.

Data

  • Historical EUR/USD minute data · HistData market-prices public
    2021-2023 · EUR/USD
    Source sentence
    We utilized a historical Forex dataset from HistData, consisting of minute-by-minute Euro/Dollar (EUR/USD) data from 2021–2023.

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • No model surpassed the ZeroR baseline accuracy of 0.5389; Logistic Regression matched it while others performed worse. negative
    Source sentence
    only Logistic Regression matched the ZeroR baseline (0.5389), while others performed worse.
  • After hyper‑parameter tuning, Random Forest, Gradient Boosting, and Logistic Regression each reached 0.5389 but still did not exceed the baseline. negative
    Source sentence
    ... improved the accuracy of Random Forest, Gradient Boosting, and kNN. However, despite these improvements, no model surpassed the ZeroR baseline of 0.5389
  • Both the MLP and LSTM neural networks achieved an accuracy of 0.5389, identical to the ZeroR model. negative
    Source sentence
    the MLP achieved an accuracy of 0.5389, identical to the ZeroR model
  • Feature selection retained only RSI features, with SMA indicators having negligible impact on accuracy. statistical
    Source sentence
    Feature selection using SelectFromModel with a Random Forest classifier retained only the four RSI features
  • Models converged to predicting the majority class, showing a failure to learn any true predictive signal. negative
    Source sentence
    most algorithms, including the LSTM and MLP, converged to a simplistic strategy of predicting the majority class, as evidenced by their confusion matrices.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 82 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (44)

  1. Machine learning for financial market prediction (2012)
  2. A survey on machine learning for stock price prediction: Algorithms and techniques (2020)
  3. Física de processos estocásticos aplicada a opções binárias no mercado financeiro (2020) — link
  4. A survey on feature selection methods (2014)
  5. Analysis of binary options trading using machine learning techniques (2023)
  6. Neural network approaches to stock price prediction in volatile markets (2022)
  7. Testing rationality in the point spread function (1988)
  8. Are financial markets really efficient? (2013)
and 36 more
  1. ESMA Decision (EU) 2018/795 of 22 May 2018 on Product Intervention Measures on Binary Options (2018)
  2. Efficient capital markets: A review of theory and empirical work (1970)
  3. New concepts in technical trading systems (1978)
  4. Multilayer feedforward networks are universal approximators (1989)
  5. Deep learning (2016)
  6. Technical Analysis for the Trading Professional (2005)
  7. Surveying stock market forecasting techniques – Part II: Soft computing methods (2009)
  8. Random forests (2001)
  9. Applied Logistic Regression (2013)
  10. Greedy function approximation: A gradient boosting machine (2001)
  11. An introduction to kernel and nearest-neighbor nonparametric regression (1992)
  12. An introduction to variable and feature selection (2003)
  13. Wrappers for feature subset selection (1997)
  14. Random search for hyper-parameter optimization (2012)
  15. Sequential model-based optimization for general algorithm configuration (2011)
  16. Practical Bayesian optimization of machine learning algorithms (2012)
  17. Deep learning (2015)
  18. Learning representations by back-propagating errors (1986)
  19. A survey of cross-validation procedures for model selection (2010)
  20. A study of cross-validation and bootstrap for accuracy estimation and model selection (1995)
  21. A systematic analysis of performance measures for classification tasks (2009)
  22. An introduction to ROC analysis (2006)
  23. Predicting Economic Trends and Stock Market Prices with Deep Learning Models (2024)
  24. Forecasting financial market structure from network features using machine learning (2021)
  25. Forecasting the NYSE composite index with technical analysis, pattern recognizer, neural network, and genetic algorithm: a case study in romantic decision support (2002)
  26. An empirical study of machine learning algorithms for stock market forecasting (2018) — link
  27. Stock market prediction system with modular neural networks (1990)
  28. Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimization (2017) — link
  29. Evaluation: From Precision, Recall and F-Measure to ROC, Informedness, Markedness and Correlation (2011)
  30. A revolução dos dados no mercado financeiro (2020) — link
  31. Nearest neighbor pattern classification (1967) — link
  32. Random decision forests (1995)
  33. Logistic Regression: Relating Patient Characteristics to Outcomes (2016) — link
  34. Stochastic gradient boosting (2002) — link
  35. Artificial neural networks: A tutorial (1996)
  36. Long short-term memory (1997)

0 of 44 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.