Research tools for design margins

MARVIN

Research tools for analysing design margins in complex engineered systems

Three integrated modules — parametric margin valuation, top-down margin allocation, and probabilistic margin risk — built around the Margin Value Method.

Module 3 · MARVIN

Probabilistic Margin Analysis

A step-by-step walkthrough of building a DSM, running CPM, and reading the combined-risk results.

Tutorial in preparation The structure below is final, but screenshots, sample DSM data, and worked example figures will be added as the module's documentation is completed.

Before you start

This tutorial walks through a full Probabilistic Margin Analysis on a simple worked example: a six-subsystem fan system, where a redesign in any subsystem may propagate to others through their interfaces.

You only need a modern web browser. The module runs entirely client-side; no installation is required.

Walkthrough

01

Add components to the DSM

Open the DSM editor and add the components of your system as rows/columns. Optionally group them into subsystems for cleaner readability and faster editing.

Screenshot: empty DSM with components added
02

Fill in direct likelihood

For each ordered pair of components, enter the probability that a change in the source will require a change in the target. Use a coarse 0–1 scale (e.g. 0.1, 0.3, 0.7, 0.9) elicited from designers' judgement — the CPM is robust to ordinal-style inputs.

Screenshot: direct likelihood matrix
03

Fill in direct impact

For each pair, enter the average severity of the propagated change — how much redesign work the target component would require if the change reached it. Same coarse scale.

Screenshot: direct impact matrix
04

Run CPM

Click the analysis button. The module composes direct probabilities along all paths between every pair of components and produces three derived matrices: combined likelihood, combined impact, and combined risk.

Screenshot: combined risk heatmap
05

Inspect per-element profiles

Click on any component to see its incoming and outgoing risk profile. Components with high outgoing risk are sources of propagation; components with high incoming risk are absorbers — both can guide where to invest in margins or in interface decoupling.

Screenshot: per-element profile chart
06

What-if exploration

Edit a direct value, decouple a pair, or insert an interface buffer, and watch the combined-risk matrix update live. Use this to test design decisions before committing to them.

Next steps