Reference scheme · Demo dataset
3c7f…a8e1Schedule risk profile
S-curve · post-mitigation primaryM3 Post-mitigation lifts the P80 by 49 days versus the un-mitigated M2 case, and the P10–P90 corridor narrows from 198 days to 142 days — the residual risk is mostly tail, not central spread.
M1 · DU-Only
P50 27 Sep 27 · P80 14 Dec 27 · P90 04 Feb 28
Duration uncertainty only — no risk register applied.
M2 · Threats Pre-mit
P50 12 Nov 27 · P80 22 Feb 28 · P90 18 Apr 28
DU + un-mitigated threat exposures.
M3 · Post-mit (primary)
P50 18 Oct 27 · P80 03 Jan 28 · P90 25 Feb 28
Mitigation actions applied — the headline post-mit position.
Top risk drivers
Tornado · Spearman rank correlation, M3 post-mitStatutory undertaker delays drive 18.4 days of mean P80 impact — three times the magnitude of the fifth-ranked driver and the single biggest residual contributor after mitigation.
Activity drivers
Tornado · activity-level rank correlationThe driver activities cluster on the western embankment and the northbound slip approach — six of the top ten activity drivers sit on the same critical path segment, so a single ground-conditions event would cascade.
Mitigation delta
Pre vs post mitigation · P80 comparisonMitigation has trimmed 49 days off the P80 — the planned action set on R-014 (Statutory undertaker delays) and R-008 (Ground conditions) accounts for 36 of those 49 days alone.
Duration uncertainty (DU) bands
DU band distribution · 138 risk-loadable activitiesThe DU profile leans Band C — 64 activities (46%) sit in the moderate band, which is consistent with a Stage 4 design maturity, but the 12 Band A activities account for 71% of the M1 P80 spread.
Band reference table (AACE 41R-08 informed)
| Band | Description | Range (min / max) | Distribution |
|---|---|---|---|
| A | Highly uncertain | −25% / +40% | Triangular |
| B | Uncertain | −15% / +35% | Triangular |
| C | Moderate | −10% / +15% | BetaPERT |
| D | Low uncertainty | −3% / +5% | BetaPERT |
Findings
SOMA QSRA Readiness V2.9 · 39 checks · 6 domainsEight findings open. Three are blockers — two DCMA leads/lags violations and one out-of-sequence cluster on the western embankment — and need closure before the next client stage gate.
| ID | Severity | Finding | Owner | Status |
|---|---|---|---|---|
DCMA-04 | Blocker | Negative leads detected on 6 relationships — A1140→A1160 (SS−5), A1220→A1240 (FS−10) and four others. Negative leads compress the critical path artificially and breach the SOMA QSRA Readiness check S3-C2. | Planner — A. Hughes | Open |
DCMA-05 | Blocker | High lag count — 14 relationships use FS lag > 10 days. Threshold is 5% of total; observed 8.6%. Recommend converting to driving activities or splitting work. | Planner — A. Hughes | Open |
SOMA-S4-C3 | Blocker | Out-of-sequence progress — 4 activities show actual start before predecessor finish on the western embankment cluster (A1080, A1100, A1120, A1140). Risk modelling integrity is compromised until resolved. | Risk Manager — D. Patel | Open |
DCMA-08 | Warning | High duration count — 31 activities have durations > 44 working days. Threshold 5%; observed 7.2%. Consider splitting long-running activities to expose progress signal. | Planner — A. Hughes | Open |
SOMA-S5-R1 | Warning | Three-point estimate sanity — R-019 (Plant availability) has min 5d / most-likely 5d / max 9d. Most-likely equals min — flag for re-estimation. | RAM — J. Sinclair | Open |
SOMA-S5-R2 | Warning | Probability boundary — R-022 (Late piling delivery) probability 100%. Risks at 0 or 100% should be moved to the schedule baseline, not modelled as probabilistic events. | RAM — J. Sinclair | Open |
SOMA-S2-C1 | Info | Correlation matrix non-PSD — original matrix had two negative eigenvalues, Higham nearest-PSD correction applied (max delta 0.07 in pairs (R-005, R-008) and (R-008, R-014)). Documented in audit log; no action required if accepted. | Risk Manager — D. Patel | Accepted |
SOMA-S1-D3 | Info | DU sample size — Band A holds 12 activities. Below the V2.9 advisory floor of 15 for monotone stability of the P80; consider re-banding two adjacent Band B activities. | Planner — A. Hughes | Accepted |
Three accepted findings (with reasons logged) are travelling with this report. The full audit trail — including HMAC-chained log entries and per-run hashes — is available on the live tool.
Per-decile risk attribution
Per-decile risk re-ranking · M3 post-mitThe P10 leaders differ from the P80 leaders — the rare, large risks (R-014 Statutory undertaker, R-008 Ground conditions) dominate the upper deciles, while the high-probability bulk-firing risks (R-005 Weather, R-019 Plant) carry the lower deciles.
Risks are re-ranked per decile using decile_factor(d) = 0.4 + 1.2 × (d / 9) — model-based, not iteration-direct (Safran iteration export doesn't preserve per-iteration risk firing; the decile weighting is the defensible approximation). Methodology stamp V2.9.
P10 leaders (low-realisation tail)
| Rank | Risk | Decile weight | Adjusted impact |
|---|---|---|---|
| 1 | R-005 — Weather contingency winter 26–27 | 0.53 | 3.0d |
| 2 | R-019 — Plant availability — paver | 0.53 | 1.1d |
| 3 | R-007 — Permits & consents — section 50 | 0.53 | 1.8d |
| 4 | R-012 — Design coordination — utilities | 0.53 | 2.2d |
| 5 | R-014 — Statutory undertaker delays | 0.53 | 9.8d |
P50 leaders (median)
| Rank | Risk | Decile weight | Adjusted impact |
|---|---|---|---|
| 1 | R-014 — Statutory undertaker delays | 1.00 | 18.4d |
| 2 | R-008 — Ground conditions M42 western emb. | 1.00 | 12.7d |
| 3 | R-003 — Land acquisition — 4 plots | 1.00 | 9.3d |
| 4 | R-021 — Buildability — earthworks Q3 | 1.00 | 7.1d |
| 5 | R-005 — Weather contingency winter 26–27 | 1.00 | 5.6d |
P80 leaders (upper-tail commitment band)
| Rank | Risk | Decile weight | Adjusted impact |
|---|---|---|---|
| 1 | R-014 — Statutory undertaker delays | 1.40 | 25.8d |
| 2 | R-008 — Ground conditions M42 western emb. | 1.40 | 17.8d |
| 3 | R-003 — Land acquisition — 4 plots | 1.40 | 13.0d |
| 4 | R-021 — Buildability — earthworks Q3 | 1.40 | 9.9d |
| 5 | R-012 — Design coordination — utilities | 1.40 | 5.9d |
P90 leaders (worst-case decile)
| Rank | Risk | Decile weight | Adjusted impact |
|---|---|---|---|
| 1 | R-014 — Statutory undertaker delays | 1.60 | 29.4d |
| 2 | R-008 — Ground conditions M42 western emb. | 1.60 | 20.3d |
| 3 | R-003 — Land acquisition — 4 plots | 1.60 | 14.9d |
| 4 | R-021 — Buildability — earthworks Q3 | 1.60 | 11.4d |
| 5 | R-008-A — Latent contamination tail event | 1.60 | 9.6d |