Paper: arXiv 2402.03953
Authors: Erdong Chen, Mengzhong Ma, Zixin Nie
Abstract
In this paper, we analyze traders’ behavior within both centralized exchanges (CEXs) and decentralized exchanges (DEXs), focusing on the volatility of Bitcoin prices and the trading activity of investors engaged in perpetual future contracts. We categorize the architecture of perpetual future exchanges into three distinct models, each exhibiting unique patterns of trader behavior in relation to trading volume, open interest, liquidation, and leverage. Our detailed examination of DEXs, especially those utilizing the Virtual Automated Market Making (VAMM) Model, uncovers a differential impact of open interest on long versus short positions. In exchanges which operate under the Oracle Pricing Model, we find that traders primarily act as price takers, with their trading actions reflecting direct responses to price movements of the underlying assets. Furthermore, our research highlights a significant propensity among less informed traders to overreact to positive news, as demonstrated by an increase in long positions. This study contributes to the understanding of market dynamics in digital asset exchanges, offering insights into the behavioral finance for future innovation of decentralized finance.
Complexity vs Empirical Score
- Math Complexity: 4.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper’s math complexity is moderate, relying on established models like Kyle’s and Shalen’s dispersion of beliefs rather than dense novel derivations. However, empirical rigor is high, featuring extensive data collection from multiple blockchain sources, daily-level frequency analysis, and direct testing of hypotheses on real trading data.
Research Flowchart
flowchart TD
A["Research Goal"] -->|"Analyze"| B["Trader Behavior on<br>CEX vs. DEX Perpetual Futures"]
B --> C{"Methodology"}
C -->|"Categorize"| D["Exchange Architectures<br>VAMM, Oracle, Hybrid"]
C -->|"Analyze"| E["Data: Volume, Open Interest,<br>Liquidation, Leverage"]
D --> F["Computational Process<br>Volatility & Correlation Analysis"]
E --> F
F --> G["Key Findings & Outcomes"]
subgraph G
direction TB
H["VAMM Model:<br>Differential impact on Long vs. Short"]
I["Oracle Model:<br>Traders act as Price Takers"]
J["Behavioral Insight:<br>Informed traders overreact to positive news"]
end
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 categorizes perpetual futures exchanges into three distinct architectural models and analyzes how each model shapes trader behavior, price formation, and market dynamics.
Source sentence
We categorize the architecture of perpetual future exchanges into three distinct models, each exhibiting unique patterns of trader behavior in relation to trading volume, open interest, liquidation, and leverage.
Assumes 2
- Pricing mechanisms for perpetual futures are designed to approximate the spot price of the underlying asset. model
Source sentence
It is paramount to recognize that the pricing mechanisms governing perpetual futures contracts are intentionally designed to approximate the spot price of the underlying asset.
- Traders respond to variations in the underlying Bitcoin price when trading perpetual futures. market
Source sentence
Consequently, traders engaged in the trading of perpetual futures are confronted with prices inherently tethered to the underlying asset’s market valuation.11 For the purpose of our analysis, and without loss of generality, we direct our focus toward an examination of how traders operating within the perpetual futures market respond to variations in the price of the underlying asset, specifically Bitcoin.
Data
- GMX trading data on Arbitrum crypto-onchain public
Source sentence
Data and Methodology GMX constructs its trading system on Arbitrum and Avalanche,3 and we focus on tradings on Arbitrum from 31 August 2021 to 26 September 2023, the data of which are retrieved from Dune (https://dune.com/queries/3110823).
- GNS trading data on Polygon crypto-onchain public
Source sentence
Similarly, GNS is based on Arbitrum and Polygon,4 and we choose the perpetual futures on Polygon from 20 December 2021 to 26 September 2023 to examine, the data of which are retrieved by our constructed queries on Dune 3 Arbitrum and Avalanche are two chains to address scalability of Ethereum (https://vitalik.eth.limo/general/2023/10/31/l2types.html), while detailed explanation can be found at: arbitrum.io and
- Perpetual Protocol V2 trading data on Ethereum mainnet crypto-onchain public
Source sentence
We also retrieve data for Perpetual Protocol V2 (based on Etherum mainnet) from 27 November 2021 to 26 September 2023 by constructing the query on Dune (https://dune.com/queries/3039245).
- Binance and Bybit daily data market-prices public
Source sentence
We retrieve data on Binance and Bybit at the daily-level frequency using API provided by Coinalyze (https://coinalyze.net/) from 05 August 2020.
- BTC daily price data market-prices public
Source sentence
Data of BTC pricing at the daily-level are collected from CryptoCompare with API (https://cryptocompare.com/).
Binance Bybit
Costs
Not applicable
Finds 4
- For LOB Model exchanges (Binance and Bybit), price volatility is positively related to trading volume but negatively related to open interest, with negative and significant coefficients for both expected and unexpected open interest. statistical
Source sentence
Coefficient estimates for expected and unexpected trading volume are positive across all the five exchanges, consistent with the findings in the previous studies on traditional futures market.5,24 For exchanges of LOB Model, i.e., Binance and Bybit, all the estimated coefficients measuring the marginal effects of expected and unexpected open interest are negative and significant, while the estimated coefficients for
- An unexpected rise in liquidation reduces market depth by lowering open interest, thereby increasing price volatility. statistical
Source sentence
This is reflected by that seven out of eight coefficients for liquidated volume in Binance and Bybit are significant and positive, with expected liquidated volume on short positions for Bybit as the exception. Thus, all else being equal, an unexpected rise in liquidation reduces market depth, via lowering open interest and liquidity. 3.2.
- In the VAMM Model (Perpetual Protocol V2), open interest impacts long and short positions asymmetrically, consistent with theoretical predictions of larger price impact when consuming liquidity. statistical
Source sentence
The asymmetrical effect of open interest on long and short positions is also predicted in the theoretical work by,3 which suggests that consuming liquidity involves a larger price impact than adding it, resulting in the larger marginal cost of a ETH buy order than the marginal benefit of a ETH sell order. Our empirical results uncover distinct variations in DEXs, particularly under the VAMM Model, where open
- Traders on Oracle Pricing Model exchanges (GMX and GNS) act as pure price takers, responding directly to underlying asset price movements. theoretical
Source sentence
In exchanges which operate under the Oracle Pricing Model, we find that traders primarily act as price takers, with their trading actions reflecting direct responses to price movements of the underlying assets. While in exchanges of LOB and VAMM Model traders’ trading activity help form the price, traders in exchanges of Oracle Pricing Model accept the price offered by the Oracle and thus act as pure price takers.
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.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (6)
- Decentralized Exchange: The Uniswap Automated Market Maker (2024) — link
- Fundamentals of Perpetual Futures — → 2212.06888
- Trading activity and price discovery in Bitcoin futures markets (2021) — link
- Delayed Reaction to Good News and the Cross‐Autocorrelation of Portfolio Returns (1996) — link
- Continuous Auctions and Insider Trading (1985) — link
- Efficient Capital Markets : A Review of Theory and Empirical Work (1994)
1 of 6 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.