Paper: arXiv 2305.07559

Authors: Christopher J. Cho, Timothy J. Norman, Manuel Nunes

Abstract

In a financial exchange, market impact is a measure of the price change of an asset following a transaction. This is an important element of market microstructure, which determines the behaviour of the market following a trade. In this paper, we first provide a discussion on the market impact observed in the BTC/USD Futures market, then we present a novel multi-agent market simulation that can follow an underlying price series, whilst maintaining the ability to reproduce the market impact observed in the market in an explainable manner. This simulation of the financial exchange allows the model to interact realistically with market participants, helping its users better estimate market slippage as well as the knock-on consequences of their market actions. In turn, it allows various stakeholders such as industrial practitioners, governments and regulators to test their market hypotheses, without deploying capital or destabilising the system.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced concepts from market microstructure theory, including the Propagator model and the Square Root law, which involves non-trivial mathematical modeling. It is empirically rigorous, as it uses real high-frequency LOB data from Binance and implements the simulation within the ABIDES platform to test market hypotheses in a controlled environment.

Research Flowchart

  flowchart TD
  subgraph G ["Research Goal"]
      G1["Model Price-Reverting Market Impact<br>in BTC/USD Futures"]
  end

  subgraph M ["Methodology"]
      M1["Multi-Agent Simulation<br>with Limit Order Book"]
  end

  subgraph D ["Data & Inputs"]
      D1["BTC/USD Futures Market Data"]
  end

  subgraph C ["Computation"]
      C1["Agent Execution &<br>Price Reversion Dynamics"]
  end

  subgraph O ["Outcomes"]
      O1["Explainable Impact Model"]
      O2["Slippage Estimation Tool"]
      O3["Risk-Free Market Testing<br>for Stakeholders"]
  end

  G --> M
  M --> D
  D --> C
  C --> O

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 presents a novel simulation framework (PRIME) that replicates key microstructural properties of the crypto market using heterogeneous agents.

Source sentence
Secondly, we present a novel simulation framework, that leverages and enhances an existing model, which produces desirable market microstructural dynamics in an explainable manner. arXiv:2305.07559v1 [q-fin.TR] 12 May 2023 Pre-print for CRC2023 Our contributions will have various implications for stakeholders in the world of market simulations.

Assumes 2

  • Data from Binance is representative of the entire cryptocurrency market for the purpose of this research. data
    Source sentence
    However, due to their size (largest exchange by transaction volume, March 2023) and the existence of high-frequency 2 Pre-print for CRC2023 traders who arbitrage between various cryptocurrency exchanges, we assume that the data we collect from Binance is representative of the entire market for the purpose of this research.
  • The noise term ϵ in the propagator model is assumed to be zero. model
    Source sentence
    Note that we ignore the noise term ϵ and assume it to be zero in our work.

Data

  • Binance USD/BTC perpetual futures Level 3 LOB data · Binance order-book public
    July 2020 to October 2021 · BTC/USD Futures
    Source sentence
    Therefore, our research is carried out on the Level 3 LOB data collected from the Binance API. More specifically, we queried data from the USD/BTC perpetual futures market from July 2020 to October 2021.

Recognised sources: Binance

Costs

Not applicable The paper reports no trading or execution result.

Finds 4

  • A positive relationship between trade size and price change is observed in the BTC/USD Futures market, though with large discrepancies due to liquidity and volatility. statistical
    Source sentence
    Looking at the resampled data from the BTC/USD Futures market (Figure 1, left image), a positive relationship between trade size and price change can be observed.
  • Impact reversion is evidenced: periods following net buy orders cause lower market impact than periods following net sell orders. statistical
    Source sentence
    This analysis indicates that periods following a previous period of net buy orders on average cause a lower market impact than those following a period of net sell orders. This means the previous order exerts market impact pressures in the opposite direction on the current period, evidencing reversion in impact.
  • The Santa‑Fe model reproduces diminishing initial market impact but fails to generate realistic order‑sign autocorrelation. performance
    Source sentence
    The diminishing initial market impact can be observed in Figure 7 and Figure 8, and the reversion in market impact can also be observed in Figures 9, 10 and 11. However, it was unable to show desirable autocorrelation in order sign to produce the idea of meta orders (Figure 12).
  • The PRIME model exhibits realistic order‑sign autocorrelation following a power‑law decay without embedding artificial stochastic processes. performance
    Source sentence
    Like the real-world data and the augmented Santa-Fe model, we observe a decay that roughly follows a power law. This is 12 Pre-print for CRC2023 especially encouraging, as the PRIME model does not embed artificial stochastic processes into the participating agents to generate this result.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 27 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.