// mostly dead, occasionally useful.
Field notes from an exhaustive, rigorous, and completely unsuccessful hunt for an edge on prediction markets. Every idea in here works. None of them makes money. You're welcome to all of it.
Thirty-six case files, catalogued by cause of death. Start with the three below.
A screenshot, a chatbot that wrote a 1,700-line warning label, and the decision to ignore it. The manifesto.
"Built to find an edge. Found the warning label was right."
A real scanner, golden-tested to the penny, found 293 arbitrages in 28,900 markets. Zero beat a savings account.
"293 opportunities. Zero that beat a savings account."
There wasn't one — just lucky coins, reward-harvesting bots, and a couple of real edges already fading.
"No lasting edge — fading sharps, vending machines, and lucky coins."
Shelfware is software that gets built and never used. Everything here is shelfware, and I'm not embarrassed about it — I'm documenting it. A well-run experiment that returns "no" is still a well-run experiment, and almost nobody publishes those.
Every claim sits on real data. Where an "edge" turned out to be a bug, I say so — because it usually was. And every post ends where its idea did: on the shelf, with a cause of death. More about the lab →
Nothing here is financial advice. I am aggressively not an advisor — a fact the results make very easy to believe.