Developer claims 74% win rate over 19 days using Claude to copy profitable prediction-market wallets
Aiden_Tech_Ai · x · 2026-09-25
A developer shared a long thread claiming an Anthropic engineer spotted him running Claude against live prediction markets at an SF cafe, and revealed his results: 214 trades, 74% win rate, +$9,437 in 19 days.
The approach: give Claude two repos — a full market archive of wallet behavior (entries, exits, timing) and a command-line scorer covering 500+ markets with no API key.
- Instead of predicting the world, the system finds wallets that consistently exit better than the crowd, isolates the pattern and only fires when the same structure reappears
- Main filter: captured value / expected value > 0.70; wallets that win but leak value on exit are ignored, while those capturing most of the move and cutting losers fast become signal
- Position sizing uses the Kelly criterion (full parameters not given)
Caveat: self-reported, unverified results.
More from coding & agent
- No BS Guide to Building an Agentic Software Factory: Faster Coding Just Moves the Bottleneck — Pavan_Belagatti · 2026-09-25
- Claude Code users urged to run '/claude-api prompt-audit' to tune configs for Opus 5.5 — JeremyNguyenPhD · 2026-09-25
- Claude Opus 5.5 builds a playable Splatoon-style game demo — Angaisb_ · 2026-09-25
- Renaming an API field is the most dangerous breaking change — diff the spec in CI — Franc0Fernand0 · 2026-09-25
- One-click narration pipeline: Claude writes shot-aware script, ElevenLabs MCP voices it — nikola_mr64990 · 2026-09-25
- Why this dev lets AI build its own tools instead of off-the-shelf products — _Stocko_ · 2026-09-25