Paper: arXiv 2308.08683

Authors: Haochen Li, Maria Polukarova, Carmine Ventre

Abstract

We take inspiration from statistical physics to develop a novel conceptual framework for the analysis of financial markets. We model the order book dynamics as a motion of particles and define the momentum measure of the system as a way to summarise and assess the state of the market. Our approach proves useful in capturing salient financial market phenomena: in particular, it helps detect the market manipulation activities called spoofing and layering. We apply our method to identify pathological order book behaviours during the flash crash of the LUNA cryptocurrency, uncovering widespread instances of spoofing and layering in the market. Furthermore, we establish that our technique outperforms the conventional Z-score-based anomaly detection method in identifying market manipulations across both LUNA and Bitcoin cryptocurrency markets.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper develops a novel, non-trivial theoretical framework by mapping order book dynamics to particle physics, involving velocity and momentum definitions, but its application is primarily focused on detecting manipulation in historical cryptocurrency data (LUNA and Bitcoin) rather than providing full backtest-ready implementation details.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Develop a physics-inspired<br>market manipulation detector"] --> B["Methodology<br>Statistical Physics Framework"]
  B --> C["Data<br>High-frequency LUNA &<br>Bitcoin order book data"]
  C --> D["Computation<br>Momentum Measure Calculation<br>Particle Motion Model"]
  D --> E{"Analysis"}
  E --> F["Flash Crash Analysis<br>Detected LUNA crash manipulation"]
  E --> G["Performance Comparison<br>Outperforms Z-score baseline"]
  F & G --> H["Key Findings<br>Novel spoofing/layering detection<br>Physics tools > traditional methods"]

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 physics‑based momentum method for analyzing limit order book dynamics to detect spoofing and layering.

Source sentence
We take inspiration from statistical physics to develop a novel conceptual framework for the analysis of financial markets. We model the order book dynamics as a motion of particles and define the momentum measure of the system as a way to summarise and assess the state of the market.

Assumes 4

  • Orders can be treated as physical particles moving on a one‑dimensional axis. model
    Source sentence
    We view the orders as physical particles with motion on a 1-dimensional axis.
  • Order size corresponds to particle mass. model
    Source sentence
    The order size corresponds to the particle mass,…
  • Active area of the order book is centred on the bid‑ask spread and follows the midprice movement. market
    Source sentence
    The active area follows the movement of the midprice: Inside this range the order submissions and cancellations are the most active, whilst outside this range the limit orders tend to keep static.
  • Synthetic spoofing orders are inserted within the passive area of the BTC/USD order book for evaluation. data
    Source sentence
    In order to further compare these two methods, we inserted synthetic order book data of traditional spoofing into the BTC/USD dataset.

Data

  • LUNA flash crash order book order-book unknown
    LUNA cryptocurrency market
    Source sentence
    We apply our method to identify pathological order book behaviours during the flash crash of the LUNA cryptocurrency, uncovering widespread instances of spoofing and layering in the market.
  • BTC/USD order book order-book unknown
    Bitcoin cryptocurrency market
    Source sentence
    our technique outperforms the conventional Z-score-based anomaly detection method in identifying market manipulations across both LUNA and Bitcoin cryptocurrency markets.
  • Synthetic spoofing dataset for BTC/USD synthetic unknown
    BTC/USD pair
    Source sentence
    In this synthetic dataset, the largest ten deviation values (rounded to two decimal places) of the momentum inside the passive area were: [60.75 52.23 23.84 23.66 22.66 20.78 20.52 20.39 18.65 17.65].

Recognised sources: Coinbase

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • The physics‑based momentum method outperforms the conventional Z‑score‑based anomaly detection in detecting spoofing for both LUNA and Bitcoin markets. performance
    Source sentence
    we establish that our technique outperforms the conventional Z-score-based anomaly detection method in identifying market manipulations across both LUNA and Bitcoin cryptocurrency markets.
  • In the synthetic BTC/USD dataset, momentum values show large deviations while Z‑score values remain almost unchanged, indicating superior detection capability of the momentum method. statistical
    Source sentence
    In contrast, the Z-score (rounded to two decimal places) of this synthetic dataset remains almost the same compared to that of the original dataset: [286.12 286.12 273.46 273.46 100.17 93.81 90.94 90.81 90.42 89.82]. We thus conclude that for the BTC/USD pair, the physics-based momentum method outperforms the tradition
  • The Z‑score method failed to identify spoofing events in the LUNA dataset and incorrectly flagged large limit orders as spoofing. negative
    Source sentence
    the Z-score method, using the same dataset, resulted in the findings presented in Table 3 that are normal orders instead of spoofing.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 10 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive (1 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.