Property Settlement Modeller for family law professionals.

s.79(4) contributions

The engine weights initial contributions on a duration-decay curve (full weight in short relationships, minimal in long), during-relationship financial contributions by relative magnitude, non-financial contributions by their durability against the asset they preserved or improved, and homemaker / parent contributions equal-weight to financial contributions per Mallet v Mallet. Post-separation contributions accrue 80–90% to the contributing party with limited dilution, reflecting the Williams v Williams line.

The engine produces a contributions percentage range, not a point estimate — reflecting the genuine analytical uncertainty in any framework application. A typical range spans 5–15 percentage points; a narrow range means the framework points firmly, a wide range means the matter is more uncertain than confident reasoning resolves.

s.75(2) future needs

The engine accumulates per-factor directional impacts: age and health (s.75(2)(a)), income / property / earning capacity (s.75(2)(b)), care of children under 18 (s.75(2)(c)), commitments (s.75(2)(d)), other-person responsibilities (s.75(2)(e)), benefits eligibility (s.75(2)(f)), and the s.75(2)(o) catch-all. The s.75(2)(naa) standard-of-living factor is contextual rather than discretely modelled.

Adjustments are presented as a percentage range that’s added to the contributions split. Future-needs is the most discretionary part of the framework; the range captures the spread two judges with identical facts would arrive at.

The just-and-equitable backstop (s.79(2))

The engine does not perform the s.79(2) test — that’s the Court’s discretionary backstop. The Modeller produces analysis; it surfaces a modelled range and explicitly flags that the test is the Court’s, not the engine’s. Beyond this point the work is human.

• No prediction of what a specific court would decide on the user’s facts
• No case-law lookup or recent-authority awareness beyond the methodology document’s publication date
• No specific-matter advice; the engine knows the framework, not your matter
• No Form 11 or Application for Consent Orders generation (the DIY Consent Orders Pack does that)
• Defined-benefit super treated at member-balance value, not actuarial; flagged in output
• Self-managed super treated as an accumulation balance; procedural complexity not modelled
• Business interests treated at user-supplied value; no goodwill, controlling-interest premiums, or related-party adjustments

The Modeller runs entirely client-side. Your inputs and your clients’ inputs do not leave the browser. We do not track Modeller usage at the matter level — no inputs, no derived values, no PDF content.

Aggregated usage telemetry (page views, completion rates per module, anonymous funnel data) is collected for product improvement. If you’d rather not contribute to that, browser ad-blockers like uBlock Origin block our analytics by default.

The optional “email me a copy” affordance on the Modeller’s result page is the only feature that touches a server. The address is used once to send the PDF, then deleted; not stored, not added to any list, not used for marketing. Most users use the Download button instead and skip server contact entirely.

Repo: github.com/pixjnr/property-settlement-modeller

License: MIT

Contribution policy: issues welcomed (especially methodology critiques from family-law specialists), external pull requests read but not merged. Methodology changes are made by the Modeller’s maintainers in response to issues, legislative change, or case-law shifts. We adopt good external work into our own commits with attribution rather than merging arbitrary contributions — the methodology has to stay coherent and accountable.

The website code, document generation, and Pack-related infrastructure remain in a separate private repository. Only the Modeller’s s.79 / s.75(2) calculation engine is open-sourced; that’s the part where transparency is the point.