← Back to projects
Prediction Market Research research

FCT Weather Bot

FCT Weather Bot is a research project that evaluates whether short-horizon weather signal can be turned into a profitable trading edge in prediction markets. The system runs entirely in paper mode and preserves a structured audit trail so each decision can be reviewed before any capital is deployed.

Last updated: January 5, 2026

Highlights

  • Quote-aware decision loop. Probability, EV, sizing logic, and explicit skip conditions are recorded for every market evaluated.
  • Paper trading first. No live capital until the structured audit log proves the signal over a meaningful settled window.
  • Audit trail is structured data, not chat output — inputs, decisions, skips, fills, and settlement results are all preserved.
  • Targeted models combined with deterministic rules. ML is not used as a black-box oracle.

Technology Stack

Python Prediction Market APIs Weather Forecast Ingestion Quote-aware Decision Loop Paper Trading Audit Log

Screenshots

What this project is

FCT Weather Bot is an internal research project. It evaluates whether public weather forecast data can be combined with prediction-market quotes to produce a profitable edge in short-horizon weather markets.

Right now the system runs only in paper-trade mode. Every decision the bot would make in live conditions is logged with full context: the input forecast, the market quote, the implied probability, the bot’s probability, the expected-value calculation, the sizing logic, and — when the market settles — the actual outcome. The point of the project at this stage is to prove the signal exists before risking real capital.

Why we publish it this way

Trading research is easy to make sound impressive. Auditable trading research is harder. FCT’s policy on bots and prediction markets is evidence-first: a research project does not graduate to live capital based on intuition, repo claims, or backtests in isolation. It graduates when the paper log shows real settled results that match the thesis over a meaningful window.

This page exists so prospective clients can see how FCT actually approaches risk — methodically, with auditability built in from day one rather than retrofitted later.

What this proves about FCT’s approach

  • Discipline: there is a clear gate between paper and production. No live money until the evidence supports it.
  • Auditability: every decision is structured data, not chat output. Inputs, decisions, skips, fills, and settlement are all preserved.
  • Practical AI: the project uses targeted models and deterministic logic together rather than treating ML as a black-box oracle.

Status

Currently in paper-trade evaluation. Public details are intentionally light. If you operate in a similar space and want to compare notes, reach out.