polymarket-ai-divergence

What This Skill Does

The polymarket-ai-divergence skill is a sophisticated algorithmic trading agent designed for the Polymarket prediction ecosystem. Its primary function is to detect pricing discrepancies between Simmer's proprietary AI consensus models—which aggregate multi-model ensemble forecasts—and the current live market price on prediction platforms like Polymarket. By identifying these gaps, the skill allows users to systematically profit from market inefficiencies.

The skill operates via a multi-stage pipeline: it scans all active markets for AI vs. market price divergence, applies strict filtering to ensure the trade meets a minimum edge threshold (defaulting to 2%) and verifying the market is zero-fee, performs risk-mitigation checks (such as detecting potential 'flip-flops' or existing position conflicts), and calculates an optimal position size using the Kelly criterion, which is then capped at a conservative 25% to protect the agent's capital. This automation ensures that trades are only executed when the statistical probability of success, as determined by the AI, significantly outweighs the market risk.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/adlai88/polymarket-ai-divergence

Once installed, verify the configuration by running python ai_divergence.py --config. You can fine-tune your trading parameters, such as the max_bet_usd or daily_budget, using the provided CLI tools or by updating the corresponding environment variables.

Use Cases

• Arbitrage Trading: Identifying and exploiting short-term mispricing in prediction markets before market equilibrium is reached.
• Portfolio Automation: Running the agent in a set-and-forget cron mode to continuously capitalize on market volatility while adhering to strict daily spending limits.
• Research & Signal Generation: Using the scan-only mode to generate trading signals based on AI consensus without automated execution, allowing for manual oversight.

Example Prompts

1. "OpenClaw, run the polymarket divergence scanner and show me all bullish markets with a divergence greater than 10%."
2. "Execute the polymarket-ai-divergence skill in live mode with a daily budget of 50 USD."
3. "List my current configuration for the AI divergence trader and set the Kelly cap to 30%."

Tips & Limitations

• Market Liquidity: Always consider the liquidity of a market before trading. Low-volume markets can have slippage that might exceed your AI edge.
• Risk Tolerance: The Kelly criterion is mathematically optimal but can lead to high volatility. The kelly_cap parameter is your most important tool for risk management; adjust it downward if you are risk-averse.
• Fees: While the tool defaults to zero-fee markets, always verify market-specific conditions, as platform rules can change rapidly. Ensure your daily_budget aligns with your overall portfolio risk management strategy to prevent excessive drawdowns during periods of high market turbulence.