Paper: arXiv 2510.22348

Authors: Aryan Ranjan

Abstract

We present a systematic trading framework that forecasts short-horizon market risk, identifies its underlying drivers, and generates alpha using a hybrid machine learning ensemble built to trade on the resulting signal. The framework integrates neural networks with tree-based voting models to predict five-day drawdowns in the S&P 500 ETF, leveraging a cross-asset feature set spanning equities, fixed income, foreign exchange, commodities, and volatility markets. Interpretable feature attribution methods reveal the key macroeconomic and microstructural factors that differentiate high-risk (crash) from benign (non-crash) weekly regimes. Empirical results show a Sharpe ratio of 2.51 and an annualized CAPM alpha of +0.28, with a market beta of 0.51, indicating that the model delivers substantial systematic alpha with limited directional exposure during the 2005–2025 backtest period. Overall, the findings underscore the effectiveness of hybrid ensemble architectures in capturing nonlinear risk dynamics and identifying interpretable, potentially causal drivers, providing a robust blueprint for machine learning-driven alpha generation in systematic trading.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 6.0/10 (capped from 9.0: transaction costs not modelled; no out-of-sample test; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor (effective empirical rigor capped from 9.0)

Why this score: The paper employs advanced statistical measures like Hurst exponents and Kullback-Leibler divergence within a hybrid ML architecture, while providing a fully backtested strategy with specific metrics (Sharpe 2.51, alpha +0.28) over a 20-year period using cross-asset data.

Research Flowchart

  flowchart TD
A["Research Goal: Forecast short-horizon market risk & generate systematic alpha using ML"] --> B["Data: Cross-asset features 2005-2025 S&P 500 ETF"]
B --> C["Methodology: Hybrid ML Ensemble<br>Neural Networks + Tree-Based Voting"]
C --> D["Process: Predict 5-day drawdowns & identify key drivers"]
D --> E["Outcomes: Sharpe 2.51, CAPM Alpha +0.28, Beta 0.51"]
E --> F["Conclusion: Interpretable, causal drivers for robust alpha generation"]

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 proposes a systematic trading framework that forecasts short‑horizon market risk and generates alpha using a hybrid machine learning ensemble of neural networks and tree‑based models.

Source sentence
We present a systematic trading framework that forecasts short-horizon market risk, identifies its underlying drivers, and generates alpha using a hybrid machine learning ensemble built to trade on the resulting signal.

Assumes 3

  • The dataset contains only 5,000–6,000 daily observations after filtering, limiting effective sample size. data
    Source sentence
    A shallow network mitigates overfitting while still capturing temporal dependencies.
  • Trading results are reported without accounting for transaction costs. execution
    Source sentence
    Note: These metrics do not currently account for transaction costs nor have they been forward tested.
  • Model performance metrics are based on in‑sample (in‑time) evaluation, not out‑of‑sample. statistical
    Source sentence
    Metrics such as ROC-AUC, precision, and recall therefore reflect in-sample performance.

Data

Backtest sample: 2005 – 2025 · S&P 500 ETF (SPY) · multi-asset · daily

  • S&P 500 ETF (SPY) daily price data market-prices unknown
    2005-2025 · S&P 500 ETF
    Source sentence
    predict five-day drawdowns in the S&P 500 ETF (SPY), leveraging a cross-asset feature set spanning equities, fixed income, foreign exchange, commodities, and volatility markets.

Costs

Not modeled A trading or execution result is reported without costs.

Source sentence
Note: These metrics do not currently account for transaction costs nor have they been forward tested.

Finds 3

Reported performance (Hybrid machine learning ensemble long/short risk‑managed strategy): Sharpe 2.51 · annual return 40.8% · max drawdown 18.1%

  • The hybrid ensemble achieves a Sharpe ratio of 2.51, an annualized CAPM alpha of +0.28, and a market beta of 0.51 over the 2005‑2025 backtest. performance
    Source sentence
    Empirical results show a Sharpe ratio of 2.51 and an annualized CAPM alpha of +0.28, with a market beta of 0.51, indicating that the model delivers substantial systematic alpha with limited directional exposure during the 2005–2025 backtest period.
  • In‑sample classification yields 0.95 overall accuracy, with class‑0 recall of 0.99 and class‑1 recall of 0.82. predictive-accuracy
    Source sentence
    Class 0 recall of 0.99 indicates that the m…
  • The strategy’s information ratio versus SPY is 1.73, maximum drawdown -18.12%, annualized return 40.84% and volatility 13.23%. performance
    Source sentence
    Information Ratio vs SPY 1.73 ... Maximum Drawdown -18.12% ... Annualized Return 40.84% ... Annualized Volatility 13.23%

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 31 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.