From mechanism
to forecast.
We build causal models of the systems that move markets — the atmosphere behind crop yields, the climate behind asset prices, the grid behind power — and license them to the firms exposed to those systems.
Models of systems,
not curves.
Most desks fit the correlation and move on. We model the mechanism underneath it — identification, structure, and the shape of the data-generating process — so the forecast survives when the regime changes.
Weather into yield, yield into price
Atmospheric and reanalysis data turned into crop-yield and food-price signals, resolved by geography, crop and phenology window.
Physical climate, repriced
Physical climate translated into asset-level impact: which cashflows move, by how much, and when the market actually reprices them.
The grid, from the inside
Interconnector capacity, merit-order economics and firm-level supply modelled into day-ahead and forward power across coupled European markets.
Shocks that travel
Weather-driven softs, metals and freight — with the energy-shock pass-through that ties otherwise separate markets together.
Built to order for the system you're exposed to. If it has a mechanism, we can model it.
Identify it. Validate it.
Deliver it.
Identify the mechanism
We start from the causal structure — what drives what, under which conditions — not from whatever correlation the last six months happen to show.
Validate on real data
Every model is tested on real, out-of-sample data. We publish the methodology and the results — including the ones that don't work.
Deliver the model
You receive a licensed, documented and monitored model — as a signal, a forecast feed, or a full pricing layer that plugs into your stack.
If you're exposed to the
system, you can price it.
Our models are built for anyone whose P&L moves with weather, climate, energy or the food chain.
We show our work.
Public, empirical notes on real data. We publish method and results — the edge stays private, the rigour is on the table.