Abstract:Harnessing the transparent blockchain user behavior data, we construct the Political Betting Leaning Score (PBLS) to measure political leanings based on betting within Web3 prediction markets. Focusing on Polymarket and starting from the 2024 U.S. Presidential Election, we synthesize behaviors over 15,000 addresses across 4,500 events and 8,500 markets, capturing the intensity and direction of their political leanings by the PBLS. We validate the PBLS through internal consistency checks and external comparisons. We uncover relationships between our PBLS and betting behaviors through over 800 features capturing various behavioral aspects. A case study of the 2022 U.S. Senate election further demonstrates the ability of our measurement while decoding the dynamic interaction between political and profitable motives. Our findings contribute to understanding decision-making in decentralized markets, enhancing the analysis of behaviors within Web3 prediction environments. The insights of this study reveal the potential of blockchain in enabling innovative, multidisciplinary studies and could inform the development of more effective online prediction markets, improve the accuracy of forecast, and help the design and optimization of platform mechanisms. The data and code for the paper are accessible at the following link: <a class="link-external link-https" href="https://github.com/anonymous" rel="external noopener nofollow">this https URL</a>.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper aims to explore and solve the following problems:
1. **Can an individual's political inclination be decoded from the betting behavior in Web3 prediction markets?**
- By analyzing users' betting behavior on decentralized finance platforms (such as Polymarket), the paper attempts to construct a quantitative index - **Political Betting Leaning Score (PBLS)** - to measure users' political understanding.
2. **How do political motives and profit - seeking motives jointly influence participants' betting strategies and market outcomes?**
- The research not only focuses on users' political inclinations but also explores the interaction between these inclinations and the pursuit of profit. Specifically, the researchers hope to reveal how these two motives jointly shape their decision - making processes by analyzing users' betting patterns in various events and markets.
### Method overview
To answer the above questions, the researchers took the following steps:
1. **Data collection and pre - processing**:
- Collected betting records of more than 15,000 user addresses on the Polymarket platform in 4,592 events and 8,742 markets, with a time span from October 2, 2020 to February 27, 2024.
- The data includes transaction hashes, timestamps, event and market names, purchase directions (buy or sell), betting targets (such as "yes" or "no", Republican or Democrat, etc.), share sizes, prices and amounts (priced in USDC).
2. **Constructing the Political Betting Leaning Score (PBLS)**:
- PBLS is calculated based on users' betting behavior, taking into account factors in multiple dimensions, such as price weight, time - decay weight, amount weight, party weight and frequency weight. The formula is as follows:
\[
\text{PBLS}_u=\frac{\sum_{i\in T_u}w_{p,i}\cdot w_{t,i}\cdot w_{amt,i}\cdot w_{pty,i}}{\sum_{i\in T_u}|w_{p,i}\cdot w_{t,i}\cdot w_{amt,i}|\cdot w_{freq,u}}
\]
where:
- \(T_u\) represents the set of transactions executed by user address \(u\).
- \(w_{p,i}\) is the price weight, reflecting the collective expectation of the market for the event outcome.
- \(w_{t,i}\) is the time - decay weight, emphasizing the importance of recent activities.
- \(w_{amt,i}\) is the amount weight, directly related to utility.
- \(w_{pty,i}\) is the party weight, reflecting support for a specific candidate.
- \(w_{freq,u}\) is the frequency weight, indicating the number of transactions in which the user participates.
3. **Validating the effectiveness of PBLS**:
- Internal validation: Ensure its consistency by comparing the PBLS distributions within different markets.
- External validation: Compare PBLS with actual public opinion poll data to verify its accuracy.
4. **Feature engineering and correlation analysis**:
- Constructed 825 features, covering users' basic information, participation status, trading behavior, success rate, profitability, preferences, time - relatedness and risk preferences, etc.
- Conducted a correlation analysis and found that 533 features are significantly correlated with PBLS (\(p < 0.05\)), providing valuable information for predicting users' political inclinations.
5. **Using machine learning to predict the PBLS of all user addresses**:
- By training the model, predict the PBLS of users who did not participate in the presidential election market, thereby expanding the scope of measuring political inclinations.
6. **Case study**:
- By analyzing the 2022 US Senate elections, further explore the relationship between political inclinations and betting behavior, especially whether high - PBLS users show higher returns when the Democrats finally win.
### Conclusions and contributions
The research constructs PB