Blog
Everything here is based on our own measured data — trade logs, price ticks and audit reports. No hypotheticals, no screenshots of other people's wins.
Disclosure: we also sell what these posts are about. Uruguabot is the Python source of the bot in the audit — the CLOB V2 execution layer with the ghost-fill fix, a dry-run simulator calibrated on measured slippage, risk rails and the accuracy monitor — $99 once, sold as-is. It ships with the record you are about to read, losses included; the full live experiment was a net loss and only the default configuration was profitable (+7.3% ROI at a 40% win rate).
-
We lost $270 building a Polymarket bot — the complete audit
August 2026 · Two bot generations, 4,000+ logged trades, every bug that cost real money, and the one configuration that worked.
-
We measured a human against 4 trading signals for 1,934 Polymarket windows
August 2026 · Momentum, mean-reversion, orderbook imbalance and an ML model vs one human's manual picks. The human won.
-
How to automate Polymarket trading in 2026 — an honest technical guide
August 2026 · CLOB V2, proxy wallets, geo-restrictions, the positions-API lag, and the slippage numbers nobody publishes.