The experience needed room for curiosity and a clear boundary around interpretation. A result should be something a person can investigate, rather than a chart that asks to be believed.
What if one assumption changed?
A useful analytical tool lets you follow that change. Inputs, model rules and results remain visible together, so the output can be understood in the context of the experiment that produced it.
Every screen.
The same intention.
The experience is reworked for a smaller screen, with the same care for the journey. Explore the live site or open it in your browser.
Follow the decision.
Question
↗Assumption
↗Model
↗Calculation
↗Interpretation
A simplified illustration of the product workflow.
How long does
a change stay?
Move the persistence parameter. The same sequence of disturbances is used each time, so the effect of the assumption stays visible.
Fixed seed. Illustrative AR(1) calculation. Separate from the authenticated geometric Brownian motion engine.
Start with a premise.
Market dynamics, agent interactions and Monte Carlo experiments ask different questions. Each model family needs its own inputs and explanation, rather than one generic simulation screen.
Make variation visible.
Parameter controls create a direct relationship between a chosen assumption and a calculated result. Repeatable examples help a reader distinguish a model change from a different random draw.
Keep interpretation close.
An output is meaningful within a model and its constraints. The experience keeps those boundaries alongside the result, giving the person using it something to reason with.
preserve the identity of each model. A mean reverting public demonstration and a geometric Brownian motion engine answer different questions and are presented accordingly.
What could we build for you? ↗
