Resolving multilateral, sequential decisions.
An option-generation engine. ROR holds it; concrete projects spin out from it.
Three layers.
Five modules.
Five feedback edges.
Each component is a pure function of its inputs plus its component state. Intermediate artefacts are cacheable, diffable, and reproducible given a fixed model version. Feedback is not optional — every upstream component subscribes to the change streams of those below.
Five modules.
Each with an interface contract.
Internal implementations vary by domain. The interfaces do not. Because modules interact only through their declared interfaces, each can be reimplemented independently — a domain with mature utility elicitation may use a sophisticated M1 and a naïve M3; a domain with heavy simulation needs will invert that.
Interest graph
Structured representation of what each actor wants, trades off, and considers infeasible. Modelled as a weighted multigraph where nodes are actors, attributes, and goals, and edges encode preferences, trade-offs, and hard constraints. Interests are preserved in their native heterogeneity — never aggregated into a social welfare function.
Option set generator
Constructs the space of feasible, executable configurations. Each option carries a feasibility witness and an explicit assumption set. Not a search returning a single optimum — a structured, filtered option space with guarantees about coverage and feasibility.
Evaluator
Scores each option against each actor's projection of the interest graph, under an explicit uncertainty model. Output is a per-actor evaluation vector with point estimates, confidence intervals, and sensitivity decompositions — never a single scalar collapsed across actors. Deterministic given a fixed (G_t, U_t, seed) triple.
Process compiler
Compiles a selected option into an executable process — a directed acyclic graph of tasks with role assignments, checkpoints, preconditions, and escalation handlers. The module emits a specification; it does not execute. The specification is handed to an execution runtime and instrumented for observability.
Governance module
Installs the incentive and commitment scheme that makes the process durable under self-interested behaviour. Emits a versioned policy object specifying opt-in thresholds, penalty functions, reward contingencies, disclosure requirements, reversibility rules. Ratification is an explicit state transition producing a verifiable commitment record.
The first options it produced.
ROR is a holding. Each option the engine produces becomes its own company; investors who resonate co-invest at that point.
Smile
The first option the engine produced. A coordination layer for what makes people smile — running, shipped at smileyou.ch.
smileyou.ch →Crew Up
A second option, queued. Shape: coordination primitives for ad-hoc small groups.
In constructioniROR
The engine itself, exposed as a reference implementation. The architecture this page describes.
In constructionCoordinated patient pathways
Cross-clinic, cross-payer decisions where each actor holds a different fragment of the situation and no shared schema exists.
Stakeholder allocation
Capital deployments with multiple principals, each with distinct utility functions, risk tolerances, and governance constraints.
Consortium projects
Multi-party infrastructure where participation, contribution, and benefit must be ratified across heterogeneous actors.
Planning across boundaries
Joint planning where the artefacts produced by each organisation's tools never share a schema, and the pipeline never closes.
None of these fields is weak on its own. The engineering gap is that they are almost never composed.
Each module resolves to existing, validated engineering and mathematical primitives. The architecture adds the composition discipline; it does not invent the primitives.
If you back option-generation engines, leave a trace.
This page is a signal test. We're measuring whether investors who fund engines — not single bets — find the shape here legible. Nothing automated happens after you submit. Christian reads every signal himself, and replies only where a real fit shows up. No deck attached, no pitch follow-up sequence.
The claim is narrow and engineering-shaped. Multilateral decisions can be engineered as systems — with explicit interfaces, versioned state, observable artefacts, defined feedback loops, reproducible runs.
This engine has produced its first option: Smile, live. The next options are queued. If the architecture resonates, the projects are where capital meets it.
Where that discipline is applied, decision throughput rises and dormant potential becomes addressable. Where it is not, no amount of capital, talent, or goodwill closes the gap, because the gap is structural.