Atropos is the Fate who cuts the thread. Prediction markets now trade tens of billions a month and desks read them as signals, but a price is not a probability: on 628 settled Kalshi markets, a contract priced around 40¢ resolved yes only 30% of the time, and a 98¢ sure thing only 84%.
We see our work as cutting that thread: a calibrated, trusted probability for each event, what it does to what you own, and the cheapest hedge that would actually have paid.
What it does
The odds on the next inflation, Fed and jobs reports from Kalshi and Polymarket, what fed-funds futures and Bitcoin options already imply, your holdings read-only from your brokerage, and every CPI and jobs release day since 2021.
How the numbers are made
Every number comes from code: an isotonic fit on settled markets, a basis against the traditional proxy, a trust score from liquidity, spread, freshness and holders, a pair screen for cointegration. Jev checks every sentence against its numbers before you read it.
Three engines
Trust the price
A prediction-market price is only a signal once it is calibrated, compared with the real market, and scored for trust.
"A contract priced near 40¢ resolved yes 30% of the time."
Price the hedge
Odds become dollars at risk for what you own, and protection is only worth buying if it paid on the days you actually lost.
"Inflation protection paid on 6 of the sample portfolio's 13 worst report days."
Pair the hedge
Sometimes the cheapest hedge is not a contract but a pair: two stocks that move against each other without being tied together.
"QQQ and VIXY: correlation −0.70, not cointegrated."
Next prints
Read from
- Kalshi
- Polymarket
- Fed funds futures
- Deribit options
Select signals
* One trade dominates the tape