Verus is the deterministic engine markst runs on. It values seven private credit asset types — direct lending, mezzanine, CRE debt, distressed, distressed bonds, illiquid public bonds and secured borrowing — and produces IFRS 13 and ASC 820 fair-value marks from a single published methodology: calibrated to the origination price, discounted on the OIS curve, and reproducible to the basis point against the engine version that produced them.

1. What Verus is

Verus is a deterministic valuation engine for private credit — not a research assistant and not an agent. Given the deal terms, the assumption set and the market inputs, it returns one mark, and the same inputs on the same engine version return that mark every time. The pricing path contains no machine learning and no stochastic draws.

Each mark is an IFRS 13 / ASC 820 fair value: classified, calibrated to a documented spread, and carried with the evidence an auditor needs to challenge it. markst holds no positions and signs no third-party opinions; you hold the book, Verus provides the engine.

2. The deterministic DCF path

The core valuation identity is deterministic:

PV = Σ_t E[CF_t] × DF_OIS(t) × exp(−s·τ)

Contractual cashflows are projected from the deal terms, adjusted for hazard-rate credit loss and LGD recovery, and discounted on the OIS curve with a facility-specific spread. There are no learned weights, no Monte Carlo drift, and no silent smoothing. A re-run from the recorded inputs reproduces the prior mark to the basis point against the engine version that produced it.

3. Day-1 calibration

IFRS 13 generally treats the transaction price as fair value on Day 1. Verus inverts that price against the DCF identity to recover an implied spread, then carries it forward with a documented market-drift adjustment at each measurement date. Every run records a calibration trace — the implied spread at origination, the drift applied, the drift source, and the final spread used — stored alongside the run.

4. Scenario blending and the sensitivity matrix

Where enterprise-value inputs are supplied, Verus reprices base, downside and recovery paths and blends them on explicit, disclosed probabilities. Each mark also carries a seven-percentile confidence range and a sensitivity matrix — the valuation repriced against its most fragile assumptions by deterministic bump-and-reprice, not a model's narrative. The result is a range and a central case, not false precision.

5. Fair-value classification (IFRS 13 / ASC 820)

Private debt is generally Level 3. Verus resolves the classification in the run output rather than asserting it by hand:

Level 2 when an observable comparable price is supplied and the DCF fair value sits within the configured tolerance of that observable.
Level 3 otherwise.

The disclosure reconciles to the engine, not to a manual label.

6. Audit lineage and reproducibility

Every deal, assumption set and run is content-addressed by a SHA-256 hash of its sorted JSON — a version id that is stable under reordering and deterministic across environments. Re-running a historical valuation from the recorded inputs reproduces the historical mark to the basis point. Deals and runs are governable objects: four-eyes approval, an explicit revaluation cadence, and an immutable calibration history.

7. What Verus does not do

No black-box ML on the mark. Generative or learned models never compute a valuation, a spread, or a recovery rate.
No silent overrides. Every figure is an input on the deal, a field on the assumption set, or a documented default. Overrides leave an explicit trace.
No undisclosed model churn. Engine changes ship under versioned methodology releases; prior runs stay reproducible against the version that produced them.

8. Coverage and reference data

Verus prices seven asset types on one engine — direct lending, mezzanine, CRE debt, distressed, distressed bonds, illiquid public bonds and secured borrowing — each with its own strategy-specific factor model. Quarterly-refreshed BDC holdings ship with the engine as a reference universe to benchmark private marks against, with a primary-source broadly-syndicated-loan corpus in build.

9. Beyond the mark

The mark is one output. The controls a quarter-end demands run on the same versioned record, so nothing is re-keyed between the valuation and the paper it ends up in:

Covenant register and monitor. Tracked covenants per deal with test history; headroom, breaches and alerts across the book.
IFRS 9. Stage 1, 2 or 3 assignment, expected credit loss per position, and attribution of the ECL move since the last snapshot.
Quarterly refresh and period close. One call revalues every carried position, re-tests covenants and re-checks staging. Periods are opened, cut off, signed off and reopened with a snapshot at each step; the NAV ledger separates re-marks from coverage change.
Committee papers. IC memo as PDF or Word, and an IC pack workbook — cover, valuation, credit analysis, risk metrics, recommendation.
Portfolio risk and stress. Spread and rate DV01 rolled up across the book, VaR, named stress scenarios, and a reporting currency pinned to Federal Reserve H.10 rates on the valuation date.
Evidence beside the mark. Companies House accounts and SEC filings read into the credit inputs, each figure carrying its source. Third-party marks recorded and compared against the Verus mark, never priced from.
Review and four-eyes. Deals and runs are submitted for approval; the reviewer must differ from the submitter. The review queue flags positions past their revaluation cadence.
API and MCP. An MCP server for analyst questions — price, stress, comparables, portfolio — and a REST API for the rest of the workflow.

10. Audit-ready by construction

A mark is audit-ready when an auditor can challenge any figure and it answers from evidence. Verus carries the calibration, the sensitivity, the classification basis, and the reproducibility on the run itself — so the answer is in the record, not reconstructed after the fact.

11. Independence you can check

The Verus mark sits beside the other marks on the deal — the valuation agent’s, the manager’s and the administrator’s — each recorded with its price per 100, the provider’s own range, its as-of date and its price basis. The record shows the gap to the Verus price in points, whether each mark falls within, above or below the Verus confidence band, and whether the Verus price sits inside the provider’s range. That is the reconciliation an auditor asks for at period end, held on the deal record rather than rebuilt in a spreadsheet. External marks are evidence only: none of them enters the valuation.

12. Where to go next

/methodology — the full published, versioned methodology.
/sample-report — a Verus valuation, screen by screen.
/model-risk — validation, challenge and change control. The methodology whitepaper is available under NDA; the SR 11-7 / SS1/23 validation pack is being assembled.
/compare/private-credit-valuation-approaches — engine versus advisory opinion versus operations platform.
/trust — security posture and disclosure.

13. Background reading

The method behind Verus, one topic at a time. Each piece stands on its own; together they are the reasoning a valuation committee or auditor will ask for.

Deterministic private credit valuation under IFRS 13 — How a deterministic IFRS 13 fair-value mark is built — a single cashflow path, a market-participant discount rate, Day-1 calibration and sensitivity disclosure.
Deterministic vs stochastic valuation — When a single-path deterministic model is right for private credit, when stochastic modelling earns its complexity, and why determinism wins on reproducibility and audit.
Day-1 calibration of private credit marks — IFRS 13 takes the origination price as fair value on Day 1. How calibration solves the implied spread, carries it forward on documented drift, and keeps the mark defensible.
How often should private credit be revalued? — Quarterly marks are a process constraint, not an accounting one. What drives revaluation frequency, why the market is moving to monthly and on-demand, and what on-demand requires.
What makes a private credit valuation audit-ready — A mark is audit-ready when an auditor can challenge any figure and it answers from evidence: calibration, sensitivity, reproducibility, independence and documentation.
Independent valuation and conflicts of interest — Why independence is the core of a defensible private credit mark, where conflicts arise across in-house, advisory and engine delivery models, and what genuine independence requires.
Private credit valuation glossary — the vocabulary, defined in plain English and in technical terms.
Case study: valuation of a unitranche loan — a worked Verus mark, input by input.

To see Verus on your own positions, request access or book a demo.

Verus — the private credit valuation engine