Rank proven crypto, prediction-market and on-chain traders, mirror their positions with one click, and audit every move against your own rules.
nutted.ai — the AI copilot from the team behind copymytrade — launches this year. Keep trading here and on theorderbook.xyz to unlock up to 20% off when it goes live.
Built-in uses a damped-trend model with an 80% confidence band (the shaded cone). For real Google TimesFM (200M-param foundation model, pretrained on ~100B points), run the included timesfm_server.py — this panel auto-detects it on 127.0.0.1:8008 and the badge flips to “TimesFM live”. Forecasts are estimates, not financial advice.
Profit comes from execution, not forecasting. Takers overpay for affirmative “YES” longshots; makers harvest that “optimism tax” just by being the counterparty. Takers post negative returns at 80 of 99 price levels.
Every Kalshi trade, 2021–2025 (CFTC-regulated). Polymarket + Kalshi data open-sourced.
Your Probability calibration & Brier score above is exactly this test: if your “YES” estimates beat the market price after fees, you’re harvesting the optimism tax instead of paying it. Favour NO at the tails, fade emotional categories (Sports, Entertainment, World Events), and demand a real edge before sizing in.
Source: Jonathan Becker, “The Microstructure of Wealth Transfer in Prediction Markets” (2026) · dataset: github.com/jon-becker/prediction-market-analysis. Figures are from the paper; shown here for education.
| Date | Instrument | Side | Size | Entry | Exit | R | P&L | Discipline | Path |
|---|
Prototype styled after theorderbook.xyz. Benchmark figures are illustrative demo data. Pre-trade verdicts and autopsies are computed locally from your own rules and trades.
The AI trading copilot from the team behind copymytrade — launching this year.
These rules drive the pre-trade verdict and discipline scoring. Stored locally in this browser.