MML

About the project

MML, Managed Mobility Lifecycle, is the operating model that The Lifecycle Transition proposes: a household contracts for a mobility outcome over a term, and the vehicle supports the contract through several successive lives instead of one sale and one ownership cycle. The series makes the argument in prose. This repository is where the argument has to survive arithmetic.

It is a computational research project, not a product. It holds the model, the assumptions with their provenance, the scenarios and the interfaces used to test the thesis; the publication interprets the model and never runs it. The two are versioned independently and cite each other by hash.

What the repository contains

• A deterministic economics core. Mobility class and requirement, the Mobility Rate composed line by line as chapter 2.3 lists it, three acquisition modes on one footing (ownership, long-term rental, MML), retention values in the chapter 4 hierarchy (mobility, component, material), the repair standard of chapter 3, the continuation decision.
• Fixtures with provenance. Every number the model reads carries its status, evidence, assumption or derived, with a source or a rationale; curves carry one status for the whole series, and the model’s own constants live in the fixtures rather than in code, so nothing is silent. The vehicle the €450 scenario rests on is documented from Toyota Motor Italia’s price list and maintenance tariff, MIMIT fuel prices, IVASS liability premiums and 77 used-market asking prices. The other six vehicles remain assumptions, and say so.
• A lifecycle graph. Requirement, shortlist, quote, contract, ownership, repair, continuation, reassignment: each an operation, connected by a persistent customer context, any of them an entry point. The automotive funnel as a graph rather than a line.
• Sensitivity and break-even. Every provenanced input shocked one at a time, ranked by what it moves; and the converse, the value of one input at which the answer changes, distance for the budget, structural life or cost of capital for parity with rental, with “none in range” recorded as an answer.
• Three assumption sets. Central, adverse and optimistic, applied by path over the same fixtures, so the publication can quote a range instead of a point.
• A CLI and an MCP server. The same operations without a language model, for reproducible runs; and as tools over the Model Context Protocol, with the Agency of chapter 6 as the orchestrator’s brief, so that any agent that speaks MCP can hold one relationship with a household across the whole lifecycle and quote only what the model returned.
• Canonical bundles. One per assumption set per model version, with figure-ready tables and a manifest of hashes, regenerated by CI from a clean checkout and compared to what is committed.

What the model says

Under model 0.2.0, central set, for Part 1’s household, two adults, two children, regular long journeys, 20,000 km a year, no home charging, €450 a month for the fixed Mobility Rate:

• €450 buys the family-touring class in its third life, not new. A six-year-old Corolla Touring Sports Hybrid with 112,000 km is €435 a month fixed, of which €18 builds the Material Capital Credit; with estimated fuel the household’s expected total is €607. The second life is €489, the new vehicle €600.
• MML as priced here costs more than long-term rental of the same vehicle, by €60 to €80 a month, with the rental’s deductibles and return settlement counted. For the gap to close on a new vehicle the structural life would have to be 23 years against the central 15, or fleet capital 1.7% against 5%. Under the optimistic set the two are at parity.
• Across the sets the third-life fixed rate spans €372 to €558. That range, not the central point, is what the series quotes.
• What moves the answer: list price, the residual curve and fuel, three inputs that are now evidence; and structural life and the consumption factor, two that are MML’s own hypotheses and can only be argued.

Every figure above is in bundles/0.2.0/ with the hash of the fixtures and of the assumption set that produced it.

Method

The model is deterministic unless a simulation explicitly introduces uncertainty. Inputs, assumptions and derived values remain distinguishable in every result. The MCP layer contains no economic logic; the publication contains no model; site builds never execute simulations. Where the chapters and the model disagreed, the chapters were the reference for definitions and the model was changed; where a published chapter was wrong, it was corrected with an erratum and a version.

Status

Model 0.2.0, September 2026. Agency and portfolio economics, Monte Carlo simulation and evidence for the remaining vehicles are the next milestones. The repository is public.