> whoami

Antonio Mora

I build AI agents for a living. In my own time, I test whether they can trade prediction markets.

Can AI agents find a trading edge?

53 ideas so far. None has beaten luck yet.

One program scored them and a second, written separately, checked the numbers, under a plan frozen first.

The audit adjusts for how many ideas I tested, so one lucky winner is unlikely to pass.

corrected for luck across all 53: BH, q = 0.10

> cat ledger.md

> ls cases

The 53 ideas by how they ended
endedideas
priced right13
needs years8
blocked7
queue5
capacity5
supply4
look-ahead2
not closed9
all53
fig. 1 The 53 ideas at the audit's snapshot (2026-10-17): 44 closed, grouped by how they ended, and 9 not closed. Outlined: the five public cases.

This site is not advice and gives no signals or trade ideas.