Paper: arXiv 2503.08692
Authors: Mahya Karbalaii
Abstract
We propose a simple yet robust unsupervised model to detect pump-and-dump events on tokens listed on the Poloniex Exchange platform. By combining threshold-based criteria with exponentially weighted moving averages (EWMA) and volatility measures, our approach effectively distinguishes genuine anomalies from minor trading fluctuations, even for tokens with low liquidity and prolonged inactivity. These characteristics present a unique challenge, as standard anomaly-detection methods often over-flag negligible volume spikes. Our framework overcomes this issue by tailoring both price and volume thresholds to the specific trading patterns observed, resulting in a model that balances high true-positive detection with minimal noise.
Complexity vs Empirical Score
- Math Complexity: 2.0/10
- Empirical Rigor: 7.5/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper uses basic statistical concepts like exponentially weighted moving averages and volatility, but lacks heavy derivations or advanced mathematics. It is heavily data-driven with a focus on real-world exchange data and practical implementation.
Research Flowchart
flowchart TD A["Research Goal"] -->|"Identify anomalous pump-and-dump events in low-liquidity crypto"| B["Data & Inputs"] B --> C["Poloniex Exchange Data<br>Price & Volume Time Series"] C --> D["Methodology: Preprocessing"] D --> E["EWMA & Volatility Calculation"] E --> F["Threshold-Based Detection<br>Adaptive Price/Volume Filters"] F --> G["Outcome"] G --> H["High True-Positive Detection<br>Minimized False Positives"] G --> I["Robust against Market Noise<br>Effective for Low Liquidity Tokens"]
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 simple yet robust unsupervised model that uses threshold‑based criteria combined with EWMA and volatility measures to detect pump‑and‑dump events on Poloniex tokens.
Source sentence
We propose a simple yet robust unsupervised model to detect pump‑and‑dump events on tokens listed on the Poloniex Exchange platform. By combining threshold‑based criteria with exponentially weighted moving averages (EWMA) and volatility measures, our approach effectively distinguishes genuine anomalies from minor tradi
Assumes 2
- A pump event must involve at least one‑third of the monthly trading volume and at least 60% of the maximum 30‑day volume. market
Source sentence
Under these conditions, we assume that in a pump event at least one‑third of the monthly trading volume occurs at the pump event
- A 20‑day EWMA combined with volatility filtering best captures true pump events while reducing false positives. model
Source sentence
In light of our findings, the adoption of a 20‑day EWMA combined with volatility filtering proves to be the most robust strategy for detecting pump‑and‑dump events in our dataset.
Data
- Poloniex token dataset market-prices public
Source sentence
Ultimately, the dataset was compiled using data from five exchanges and 450 symbols, amounting to 480 hourly candles per trading pair.
Costs
Not applicable
Finds 2
- The best performance was achieved with a 90% price increase relative to the 12‑hour moving average and a 400% volume increase, yielding the highest F1 score of 0.84 when using EWMA combined with volatility. performance
Source sentence
As observed, the best performance on our dataset was achieved with a 90% increase in the High price—relative to the 12‑hour moving average
- Setting No. 4 (price threshold 80% and volume threshold 300%) demonstrated the best overall true‑positive detection among the five tested configurations. performance
Source sentence
Setting No. 4 demonstrates the best performance.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 11 related papers). Coverage is partial, so this is neither replication nor refutation.
1 paper in the archive
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.
Cited References (24)
- Cryptocurrency pump and dump schemes: Quantification and detection. (2019)
- To the moon: Analyzing cryptocurrency pump-and-dump schemes. (2018)
- The anatomy of a pump-and-dump scheme. (2019)
- The 2024 Crypto Crime Report. (2024)
- Guidance for a Risk-Based Approach to Virtual Assets and Virtual Asset Service Providers. (2019)
- Securities Exchange Act of 1934. (1934)
- Cryptocurrency Pump and Dump Schemes: Quantification and Detection (2023)
- Cryptocurrency Pump and Dump Schemes: Quantification and Detection (2019)
and 16 more
- Hybrid Cryptocurrency Pump and Dump Detection (2020)
- Machine Learning-Based Detection of Pump-and-Dump Schemes in Real-Time (2024) — link
- Sequence-Based Target Coin Prediction for Cryptocurrency Pump-and-Dump (2023)
- Pump-and-Dump Schemes in Financial Markets: Evidence and Analysis (2008)
- Pump, Dump, and then What? The Long-Term Impact of Cryptocurrency Pump-and-Dump Schemes (2023) — link
- Fraud, Regulation and Enforcement: An International Perspective on Market Manipulation (2015)
- Information Asymmetry and Fraudulent Trading: Evidence from Pump-and-Dump Schemes (2010)
- Exploratory Data Analysis (EDA) for Banking and Finance: Unveiling Insights and Patterns (2024)
- Major Issues in High-Frequency Financial Data Analysis: A Survey of Solutions (2025)
- A Historical Analysis of Pump and Dump Schemes and Market Manipulation (1999)
- Price Manipulation in the Bitcoin Ecosystem (2018)
- Candlestick Patterns and their Role in Market Data Analysis (2012)
- Deep Learning for Anomaly Detection: A Survey (2019) — link
- Unsupervised Anomaly Detection in Time Series: Challenges and Opportunities (2020)
- Anomaly Detection: A Survey (2009)
- Conditional Anomaly Detection in Complex Data: Methods and Applications (2019)
0 of 24 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.