What holds up
when the world
changes?
A diversified portfolio. A confident forecast. A clever hedge. Each depends on assumptions. Let’s make them visible—and see what happens when they move.
Thirteen investigations. Three hands-on stories. Evidence you can inspect.
Measure the exposure
What can lose value, by how much, and over what horizon?
Challenge the assumptions
What changes when markets move together or forecasts miss?
Evaluate a response
Does a different decision reduce the risk after its costs?
Start here · no formulas required
Three ways to question a model.
Follow the explanation, change an assumption, read the evidence.
The connection problem
When diversification fails
Stress the assumptions behind your portfolio. See what happens to losses when volatility rises and assets move together.
Follow the story
The confidence problem
Can we trust the forecast?
A confidence level is a promise. Test how well prediction intervals keep it when the data changes.
Follow the story
The decision problem
Does learned hedging help?
Put simple and learned strategies on the same paths. Explore the price of protection and the impact of trading costs.
Follow the story
The research atlas
One question leads to another.
Market & Tail Risk
What could we lose, together?
5 studies02Credit & Contagion
Whose failure becomes our loss?
3 studies03Stress & Climate
What if the world changes?
2 studies04Model Reliability
Does the forecast keep its promise?
1 studies05Hedging & Allocation
Which response earns its complexity?
2 studiesA useful result can be an uncomfortable one
More complexity has
to earn its place.
In the saved benchmarks, learned hedging, graph learning, and reinforcement learning each lose to a simpler baseline on the reported risk metric. That is part of the research—not a result to hide.
How we make comparisons fair