We don't run the markets we grade. We built our own predictive engine
a world model that reads the future from incentives & physics.
This is how it thinks.
A world model that forecasts mass human behaviour — elections, policy, adoption, conflict, markets — by reading the incentive structures that actually drive how crowds act, fused with the physics of how systems tip and Tetlock-style calibration. Ask it about the future and it shows its entire reasoning: the actors, their real incentives, the variables that govern the outcome, and a probability with its work.
Map the real — including hidden and perverse — incentives of every actor, and how people behaved under similar structures historically. People respond to incentives; read them and you predict the crowd, especially the counter-intuitive cases.
Treat the actors as a coupled system — allies pull together, rivals pull apart — and solve for the collective outcome the way physics solves a field of interacting particles. It surfaces coalitions, cascades, and genuine deadlock.
Before a system flips, it loses resilience — and that leaves a statistical fingerprint. The engine watches for it, so it can flag when a market or a regime is closer to breaking than it looks.
Reason the question several independent ways, pool them the mathematically-correct way, get confident only as far as they actually agree — and anchor to the historical base rate and the crowd instead of inventing certainty.
These are the ideas. The lenses, the solver, and the prompts that make them work are ours — and the model's knowledge grows every day. Browse the live Knowledge Core →
A prediction is only as honest as its terms. So before it forecasts, the engine writes the falsifiable resolution itself — the exact thresholds and the deadline that decide YES vs NO — and tells you what it assumed. Every call is then graded in public as it resolves. No cherry-picking, no hiding the misses.
Point the engine at any existing prediction market and get its independent second opinion — our number vs theirs, plus whether the market's terms even hold up.
Ask any question about the future — even one no market lists. The engine defines the terms, runs the model, and calls it.
WICK members get the Engine — free.
The Almanac + the ratings are open to all. The Engine is the members' edge.