Paper: arXiv 2309.02994

Authors: Eyal Neuman, Wolfgang Stockinger, Yufei Zhang

Abstract

We consider an offline learning problem for an agent who first estimates an unknown price impact kernel from a static dataset, and then designs strategies to liquidate a risky asset while creating transient price impact. We propose a novel approach for a nonparametric estimation of the propagator from a dataset containing correlated price trajectories, trading signals and metaorders. We quantify the accuracy of the estimated propagator using a metric which depends explicitly on the dataset. We show that a trader who tries to minimise her execution costs by using a greedy strategy purely based on the estimated propagator will encounter suboptimality due to so-called spurious correlation between the trading strategy and the estimator and due to intrinsic uncertainty resulting from a biased cost functional. By adopting an offline reinforcement learning approach, we introduce a pessimistic loss functional taking the uncertainty of the estimated propagator into account, with an optimiser which eliminates the spurious correlation, and derive an asymptotically optimal bound on the execution costs even without precise information on the true propagator. Numerical experiments are included to demonstrate the effectiveness of the proposed propagator estimator and the pessimistic trading strategy.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper presents highly advanced mathematical frameworks including stochastic control, Volterra equations, nonparametric estimation, and regret analysis with heavy derivations and theoretical proofs. Empirical validation is limited to included numerical experiments demonstrating effectiveness, but lacks real-world data backtesting or code implementation details.

Research Flowchart

  flowchart TD
  Start(["Research Goal"]) -->|"Estimate Price Impact Kernel & Design Optimal Liquidation Strategy"| Inputs["Data/Inputs<br>Static Dataset<br>- Correlated Price Trajectories<br>- Trading Signals<br>- Metaorders"]
  
  Inputs -->|"Nonparametric Estimation"| Method1["Key Methodology<br>Propagator Estimator<br>Quantifies Accuracy vs Dataset"]
  Method1 -->|"Quantifies Bias & Uncertainty"| Prob["Identify Suboptimality Causes<br>1. Spurious Correlation<br>2. Biased Cost Functional"]
  
  Prob -->|"Offline RL Approach"| Method2["Key Methodology<br>Pessimistic Loss Functional<br>Optimiser removes Spurious Correlation"]
  Method2 -->|"Computational Process"| Comp["Optimisation<br>Minimise Execution Costs<br>Account for Uncertainty"]
  
  Comp --> Outcomes["Key Findings/Outcomes<br>1. Accurate Propagator Estimation<br>2. Asymptotically Optimal Bound<br>3. Pessimistic Strategy outperforms Greedy"]
  
  Start -->|"Strategies to liquidate risky asset"| Outcomes

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 an offline reinforcement learning approach with a pessimistic loss functional that accounts for uncertainty in the estimated price impact propagator and eliminates spurious correlation, yielding an asymptotically optimal bound on execution costs.

Source sentence
By adopting an offline reinforcement learning approach, we introduce a pessimistic loss functional taking the uncertainty of the estimated propagator into account, with an optimiser which eliminates the spurious correlation, and derive an asymptotically optimal bound on the execution costs even without precise information on the true propagator.

Assumes 6

  • The noise process ε(n) is conditionally sub‑Gaussian. statistical
    Source sentence
    In the following assumption we classify the noise in (2.11) as conditionally sub-Gaussian. Assumption 2.8. The agent has an offline dataset D of size N ∈N as in Definition 2.4 and there exists a known constant R > 0 such that for all n = 1, . . . , N, EP′ h exp  ⟨v, ε(n)⟩  | Gn−1 i ≤exp R2∥v∥2 2  , for all v ∈RM.
  • A static dataset D containing N price and signal trajectories and the corresponding metaorders exists. data
    Source sentence
    In the 2 following, we assume that there exists a dataset D with N price and signal trajectories, including the corresponding agents’ metaorders following the dynamics in (1.1), D = n (S(n), u(n), A(n)) | n = 1, . . . , N o , (1.2) see Definition 2.4 for details.
  • The price impact kernel G (propagator) exists and determines execution prices via the linear relation in (1.1). model
    Source sentence
    Propagator models assume that there exists a matrix G = (Gi,j), known as the propagator or the price impact kernel, such that the asset’s execution price S is given by Sti+1 = St1 + i X j=1 Gi,jutj + Ati + εti, i = 1, . . . , M, (1.1) where the process A + ε represents the fundamental (or unaffected) asset price.
  • The dataset may contain any price‑adaptive trading strategies u(n), not only a single deterministic strategy. data
    Source sentence
    (ii) We allow for any price adaptive strategies u(n) in the dataset D, in contrast to choosing only one deterministic strategy that will repeat itself (i.e., u(n) = u(m) for all 0 ≤m, n ≤N) as in Theorem 2.10 of [34] .
Show 2 more
  • Non‑convolution propagators are allowed; when the propagator is convolutional, faster convergence rates are obtained. model
    Source sentence
    (iii) We allow for estimation of non-convolution propagators G and we show that in the convolution case the convergence rate is substantially faster (see Remark 2.15).
  • Synthetic data generated as described (fixed trading days, 5‑minute bins, specific signal and price dynamics) is representative for evaluating the estimators and strategies. other
    Source sentence
    3 Numerical Illustration In this section, we examine the performance of the propagator estimators in Section 2.2 and the pessimistic strategies presented in Section 2.4. We start by describing the construction of the synthetic dataset D for our experiments.

Data

  • Synthetic dataset D synthetic unknown
    Source sentence
    In the 2 following, we assume that there exists a dataset D with N price and signal trajectories, including the corresponding agents’ metaorders following the dynamics in (1.1), D = n (S(n), u(n), A(n)) | n = 1, . . . , N o , (1.2) see Definition 2.4 for details. 3 Numerical Illustration In this section, we examine the performance of the propagator estimators in Section 2.2 and the pessimistic strategies presented in

Costs

Modeled Costs are part of the reported headline results.

Execution costs are modeled via the propagator kernel G and the cost functional; the greedy strategy based on the estimated propagator incurs suboptimal costs, while the pessimistic strategy reduces these costs.

Source sentence
We show that a trader who tries to minimise her execution costs by using a greedy strategy purely based on the estimated propagator will encounter suboptimality due to so-called spurious correlation between the trading strategy and the estimator and due to intrinsic uncertainty resulting from a biased cost functional. The pessimistic strategy takes the model uncertainty into account and achieves more stable

Finds 5

  • A pessimistic offline learning strategy eliminates spurious correlation and achieves an asymptotically optimal bound on execution costs even without precise knowledge of the true propagator. theoretical
    Source sentence
    By adopting an offline reinforcement learning approach, we introduce a pessimistic loss functional taking the uncertainty of the estimated propagator into account, with an optimiser which eliminates the spurious correlation, and derive an asymptotically optimal bound on the execution costs even without precise information on the true propagator.
  • Numerical experiments demonstrate that the proposed non‑parametric propagator estimator and the pessimistic trading strategy are effective. performance
    Source sentence
    Numerical experiments are included to demonstrate the effectiveness of the proposed propagator estimator and the pessimistic trading strategy. The pessimistic strategy takes the model uncertainty into account and achieves more stable performance and drastically reduces the execution costs.
  • Direct Volterra (non‑convolution) estimation yields large errors unless the dataset contains sufficiently noisy trading strategies. statistical
    Source sentence
    Using a synthetic dataset, we illustrate the following characteristics of our methods: 17 • Directly estimating a Volterra propagator using (2.16) may result in large estimation errors unless the dataset contains sufficiently noisy trading strategies. Our numerical results show that in the present setting, the Volterra estimator (2.18) yields large estimation errors, as the observed trading speeds in the dataset D are
  • Imposing a convolution structure and shape constraints dramatically improves estimation accuracy, achieving relative errors of order 10⁻³ with modest sample sizes. statistical figure from elsewhere in the paper
    Source sentence
    By imposing a convolution structure and shape constraints of the estimated model, (2.25) significantly improves the esti-mation accuracy, even with a smaller sample size. • Minimising the execution costs in (2.4) after substituting the estimated propagator instead of the true one, yields a greedy strategy that is very sensitive to the accuracy of the estimated model and also creates unfavorable transaction costs. In
  • The pessimistic strategy, which accounts for model uncertainty, provides more stable performance and drastically reduces execution costs compared to a naive greedy strategy. performance
    Source sentence
    The pessimistic strategy takes the model uncertainty into account and achieves more stable performance and drastically reduces the execution costs. Our numerical results show that the performance of a naive greedy strategy using the estimated model is very sensitive to the quality of the estimated model.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 17 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 2 papers in the archive (2 by independent authors)

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.