Write the thesis
Start with the market call in plain language. No dashboard archaeology, no vague alpha thread.
campaign · @evapredicts
agent market calls need receipts.
Prediction agents are about to flood the timeline with confident takes. Eva is the layer that makes those takes inspectable: cited signals, visible revisions, and a public author record.
campaign hypothesis
trust shifts from predictions to records.
The winning prediction surface will not be the one with the loudest calls. It will be the one where every call can show what changed, when, and why.
- S1 · source — market / fact / effect
- S2 · revision — visible history
- S3 · author — X + wallet trail
why now
public AI forecasts are useless without an audit trail.
A screenshot of odds is not a thesis. A thread with no cited signal is not a record. Eva turns the prediction into a living object readers can inspect before they trust it.
Attach the evidence
Prediction markets, facts, second-order effects, and source links sit inside the same object.
Keep the revision trail
When odds move or facts change, the update becomes part of the record instead of replacing the take.
target audience
for people whose market calls need to be checked later.
prediction-market writers who want their reasoning to survive the timeline
Give readers the object behind the take, then let @evapredicts distribute the clean version.
crypto analysts turning broad theses into inspectable artifacts
Give readers the object behind the take, then let @evapredicts distribute the clean version.
agent builders who need receipts before automating public market commentary
Give readers the object behind the take, then let @evapredicts distribute the clean version.
metric to watch
does the receipt framing create intent?
Watch sessions and CTA clicks with utm_campaign=agent_receipts, then compare follow clicks, example-thesis clicks, and compose starts against the SpaceX launch thesis campaign.