| name | Prediction Market Weather Trading |
| description | Production weather market trading system extracted from alteregoeth-ai/weatherbot — airport station resolution matching eliminates 3-8F systematic error, multi-source forecast selection (HRRR/ECMWF/METAR) weighted by geography and time horizon, self-calibrating per-city sigma replaces static probability assumptions after 30+ samples, quarter-Kelly with hard cap sizes every position defensively, and a three-exit framework (stop-loss + trailing + forecast-change) manages risk on price, profit, and information dimensions simultaneously. |
| version | 1.0 |
| format | completion-engine |
| workflows | 3 |
| source | MES 3.0 Deep Extraction — alteregoeth-ai/weatherbot (2,009 lines production Python) |
Prediction Market Weather Trading
The bot wins because it solves a coordination problem most traders ignore — weather markets resolve on specific airport stations, not cities. Using KLGA coordinates instead of "New York City" coordinates eliminates 3-8 degrees F of systematic error on markets with 1-2 degree F buckets. Combined with multi-source forecast cross-validation and learned per-city sigma calibration, this creates persistent informational edge against traders using generic weather data.
Core Principle: More accurate data at the resolution point beats better trading at the wrong location.
Available Workflows
| # | Workflow | Produces | Use When |
|---|
| forecast | Market Forecast & Edge Detection | Ranked opportunity table with probability, EV, Kelly sizing, and top-3 actionable edges | Scanning active Polymarket weather markets for mispriced positions |
| execution | Trade Execution Plan | Complete trade ticket with entry, position size, all five exit scenarios, and decision tree | You have an identified edge and need the full lifecycle plan before placing a trade |
| calibration | Self-Calibration Review | Per-city sigma report, parameter optimization recommendations, and updated config.json | After 30+ resolved markets, reviewing model accuracy and tuning parameters |
Quick Reference
- Genius Context: genius.md — load before any workflow
- Source Extraction: weatherbot-extraction.md
- Reference Implementation: alteregoeth-ai/weatherbot (GitHub) — bot_v1.py (450 lines) + bot_v2.py (1,050 lines)
- Key Constraint: ALWAYS map markets to airport ICAO codes before forecasting. Never use city-center coordinates.
- Architecture: LLM reasons, deterministic code executes. Separation is absolute and non-negotiable.
Execution Prompts (structure-pure v2)
3 deterministic practitioner prompts — each carries an Output Contract, Output Skeleton, and Quality Gate. When a deliverable matches one, Read it and honor its contract instead of improvising the output shape.
- Weatherbot System — Market Scan & Edge Detection —
skills/prediction-market-weather-trading/references/prompts-v2/market-scan-edge-detection.md
- Weatherbot System — Self-Calibration Review —
skills/prediction-market-weather-trading/references/prompts-v2/self-calibration-review.md
- Weatherbot System — Trade Execution Blueprint —
skills/prediction-market-weather-trading/references/prompts-v2/trade-execution-blueprint.md
