Nice release. At ~$5k realized the dollar number says very little either way; the informative stats are fill count and average edge per fill relative to the price paid. Maker edges tend to be small and frequent, so they cross the 95% line on surprisingly modest samples, while directional records at similar P&L almost never do. If you ever publish the fill log I'd be curious to run the significance math on it.

Author here, yhis is the third post in a series - the retro and the post-mortem are the story of how it made ~$5k and then bled a chunk back. If you'd like to read through the brief article I have for this release, feel free to have a look: https://kacho.io/open-sourcing-my-polymarket-bot

Honest framing so nobody feels short-changed - it's the retired Python version (I run a Rust rewrite now), and the scrapers that kept my odds fresh aren't in it - that plus speed was the edge, and it's why giving the code away costs me nothing. It's a working foundation, not a money printer. The analytics stack I built to debug it is in there too, and those run against any public wallet.