aaaaqwq/simmer-settlement-farming

Settlement Farmer

Buy near-certain outcomes at 95‑99¢ during Oracle‑resolution lag.

When to Use This Skill

Use this skill when you want to:

• Capture risk‑free profit after market resolution but before Oracle finalization
• Buy high‑probability (>95%) outcomes at a discount
• Automate settlement farming (buying at 95‑99¢, holding until $1.00 settlement)
• Monitor resolved markets for slow price adjustment

When NOT to Use This Skill

• Markets haven't resolved yet (no settlement farming opportunity)
• Prices are already at $0.99+ (no discount available)
• Oracle has already finalized (market settled, no opportunity)
• User wants high returns quickly (settlement farming is 1‑5% return, not 50%)
• Trading in sandbox mode (no resolved markets available in sandbox)
• No Gamma API access (resolved flag detection degraded)

Core Strategy

1. Scan active markets for high implied probability (>95%) but price < $0.99
2. Identify resolved markets (via Gamma API resolved flag) or near‑expiry markets
3. Place limit orders on the winning side at a slight premium to current bid
4. Hold until settlement – collect $1.00 per share (risk‑free after resolution)

Installation & Setup

1. Ensure Simmer SDK is installed

pip install simmer-markets

2. Set environment variable

export SIMMER_API_KEY="your_api_key_here"

3. (Optional) Configure settings

Create config.json in the skill directory:

{
  "entry_threshold": 0.95,
  "max_price": 0.99,
  "min_order_size": 1.0,
  "max_order_size": 10.0,
  "max_positions": 3,
  "use_gamma": true,
  "gamma_api_key": ""
}

Usage

Dry‑run (scan only)

python settlement_farmer.py

Execute in sandbox

python settlement_farmer.py --live --sandbox

Real‑money trading (requires confirmation)

python settlement_farmer.py --live --real

Command‑line options

• --live – execute trades (otherwise dry‑run)
• --sandbox – use Simmer sandbox (default)
• --real – use real trading environment
• --dry-run – scan only, no trades (default)
• --limit 50 – scan up to 50 markets
• --config /path/config.json – custom config file

Algorithm Details

Opportunity Detection

1. Fetch active markets from Simmer (client.get_markets())
2. For each market, retrieve order book (client.get_order_book())
3. Compute mid‑price for YES and NO tokens
4. Select side with higher implied probability
5. Filter:
   • Probability ≥ entry_threshold (default 0.95)
   • Price ≤ max_price (default 0.99)
   • Market resolved or near expiry (if Gamma API enabled)
6. Calculate edge: 1.0 - price

Order Execution

• Side: BUY for YES, SELL for NO (shorting the losing side)
• Order type: Limit order at current mid‑price (or slightly above)
• Size: Between min_order_size and max_order_size USD
• Source tag: sdk:settlement‑farming for portfolio tracking

Risk Management

• Maximum max_positions concurrent positions
• Only trade in sandbox until strategy validated
• Monitor fill rates and slippage

Integration with OpenClaw Cron

To run settlement farming automatically, create a cron job:

{
  "name": "settlement‑farmer‑hourly",
  "schedule": {
    "kind": "cron",
    "expr": "0 * * * *",
    "tz": "Asia/Singapore"
  },
  "payload": {
    "kind": "agentTurn",
    "message": "Run settlement farmer dry‑run for 50 markets and report opportunities.",
    "model": "deepseek/deepseek‑reasoner",
    "thinking": "low"
  },
  "sessionTarget": "isolated",
  "enabled": true
}

Limitations & Known Issues

• Token mapping: The script currently assumes YES token is first in order book; may need adjustment for correct token‑outcome mapping.
• Gamma API dependency: Optional but recommended for resolved flag; public endpoint may rate‑limit.
• Sandbox markets: Limited availability of resolved markets in sandbox; test with real environment carefully.

Performance Expectations

• Win rate: ≈100% (if market is correctly resolved)
• Return per trade: 1‑5% (price discount)
• Holding period: Minutes to hours (until Oracle finalizes)
• Capital efficiency: High (low risk, modest returns)

Future Enhancements

1. Direct CLOB integration – bypass Simmer for lower latency
2. Multi‑market portfolio optimization – allocate capital across multiple opportunities
3. Oracle‑finalization monitoring – auto‑exit when price reaches $0.999
4. Cross‑platform settlement farming – extend to other prediction markets (Kalshi, Metaculus)

---

📊 Strategy validated: Settlement farming extracted ~$40M from Polymarket in 2024‑2025 (academic research). Window narrowing but still viable in 2026 for automated traders.