Traceable quantitative schedule risk analysis.
From input validation to traceable quantitative cross-checks — alongside your approved engine of record.
Four phases. Same data.
Validate
Four input lanes converge on one canonical schema. Quality checks point to the source row.
Cross-check
The internal deterministic Monte Carlo engine produces independent cross-check evidence for comparison.
Review
S‑curves, drivers, sensitivity and mitigation deltas are reviewed within the non-reliance boundary.
Trace
Outputs are designed to trace to source and retain run provenance for supervised audit replay.
Quality checks that run as your team works.
No more "let's run the validator at the end and find out everything's wrong." Risk Studio validates as data comes in.
- Four input lanes — manual entry, paste‑in, standard template, non‑standard importer.
- 17 canonical columns. Full or Split impact is mandatory, with inline help on every field.
- Distribution validation — Triangular, BetaPERT, Uniform, Discrete, Lognormal and Constant. Min ≤ ML ≤ Max enforced where applicable.
- Correlation matrix checks — symmetry, diagonal = 1, values in [-1, 1], positive semi‑definite via Higham PSD.
- Risk‑to‑activity mapping is checked end to end: orphan risks flagged, split weights summed, single and multi mappings validated.
Deterministic internal cross-check evidence.
Higham PSD, NORTA and six supported distributions. Formal independent numerical review remains open, so results are published as cross-check evidence.
- Layered model approach — duration uncertainty and risk layers are kept explicit for comparison and traceability.
- Absolute-days duration uncertainty — a workshop mode that captures duration uncertainty in absolute days rather than percentages, so estimators work in the units they defend.
- Higham PSD repair guarantees positive semi‑definite correlation matrices, so sampling stays stable on analyst‑built matrices.
- NORTA sampling — correlated random variates with target marginals preserved.
- Six distributions — Triangular, BetaPERT, Uniform, Discrete, Lognormal and Constant.
- Spearman rank attribution drives the tornado. It's the right measure for skewed schedule outputs and what the AACE guidance recommends.
- Per‑decile attribution — which risks drive the P10 vs the P80. Honest about model granularity.
One run. Every briefing. Same numbers.
Review the controlled workflow with reference data.
We'll explain the validation, traceability and non-reliance boundary. Do not send client files through the public site.