Loading…
A quant terminal for prediction markets. We model fair value on every Kalshi contract we cover, surface the edge, and tell you how strongly to take it. No tips. No hype. Just numbers.
14-day trial · no card required · cancel anytime
Most of the time the market gets the probability right. A few times a day, it doesn’t. The gap is the edge — and that’s all we hunt.
EXAMPLE — One historical Fed-rate trade, walked through.
Buy YES at $0.58. If the Fed cuts 25bps, you get $1.00. If they don't, you get $0. Same in reverse for NO.
A $0.58 price means the market thinks there's a 58% chance the Fed cuts. Most of the time, this is right.
Our model puts fair value at $0.62 (62% chance the cut lands). Market is at $0.58 — underpriced by 4¢. Buy YES.
Running average P&L across 1,000 simulated trades of this same setup. Wild at first — single trades pay +$0.42 or −$0.58. By trade 1,000, it’s converged on the fair-value spread of +4¢ per share. That’s the edge that compounds.
No. Alpha Geist surfaces fair-value estimates and the edge versus market price. You decide whether to act and execute on the venue yourself. The platform never touches your brokerage or places orders.
We model resolution probability from historical data and surface where the market price diverges from it. It's analytics, not tips — and explicitly not investment advice.
Kalshi and Polymarket event contracts. Mentions and FOMC categories first, with Sports as the next expansion. Coverage grows as the model is validated on each category.
A model trained on resolved outcomes — what actually happened — rather than intraday price momentum. The methodology page walks through the feature inputs and confidence intervals in detail.
There is a free tier with a limited signals view, plus paid tiers that unlock the full signal set, backtest access, and (on the top tier) programmatic API access.
The live, publicly tracked record begins Q3 2026. Until then the signal card and figures shown on this page are illustrative samples, not live data.