Worked examples

Worked examples

Four models that ship with Quantik, from a half-million-dollar council road job to a $54M highway set up for four funding authorities. Every number below comes from a Quantik run: 10,000 iterations, Latin Hypercube sampling, seed 42.

01 · Bayview road upgrade

File: examples\Example QRA - Bayview.xlsx

A council road upgrade: site clearing, earthworks, pavement and surfacing, drainage, concrete, driveways and provisional items. It is built the way most first QRAs are, and it is the workbook used in your first run.

The model

• Inherent sheet: nine estimate sections with a base cost and a percentage range each, as =SYN.PERT(C2*(1+D2), C2*(1+E2), C2*(1+F2), SYN.NAME(B2)).
• Risks sheet: eight contingent risks as SYN.BERNOULLI(p) × SYN.PERT(min, ml, max). Seven are "Possible" (37.5%); tree removal is "Unlikely" (17.5%).
• Opportunities sheet: two opportunities, a competitive tender market and a good cut/fill balance.
• Summary sheet: =SYN.OUTPUT("Total project cost", inherent + risks − opportunities).

The result, and the lesson

The same model was run with its ranges read two ways: as hard minimum and maximum (the SYN.PERT workbook) and as P10/P90 (the same model written with RiskPertAlt, as Cenex's @RISK workbooks were).

Same ranges, same risks: a 44% larger P90 contingency purely from how the ranges are read. Decide it deliberately, and use SYN.RANGE so the decision is one setting.

This model was also used to prove the add-in: run through Excel and through Quantik's engine, every one of the 10,000 iterations matched to floating-point rounding.

02 · The BOQ risk model

File: examples\Example QRA - BOQ Risk Model.xlsx

A complete, governed QRA built on a bill of quantities. It is the template the authority versions are built from, and the best place to see every Quantik feature in context.

Try: run it, then open Contingency, Sensitivity and Browse Results. Switch Risk Model B8 between Residual and Unmitigated, and compare the two with Scenarios.

03 · Detailed highway job

A regional highway upgrade, generated to exercise Quantik at realistic scale and reviewed page by page in a 27-page QRA review pack.

03.1 · The line-item baseline

Run as line-item ranging (137 uncertain inputs, ranges as min/max, ρ = 0.3 within each BOQ section, sections independent), the client report reads:

The base sitting at P3 is expected, because every Range Basis default is skewed upwards. But a P50 → P90 spread of only 3.3% on a business-case estimate is narrow, and the Estimate Maturity sheet says so. That prompted the method work below.

03.2 · What moves the P90

First, the 0.6 review: sections independent of each other, and risk impacts read as min/max whatever the setting.

Correlation within sections barely matters, because the eight sections still average each other out. Quantik 0.7 adds correlation between sections and the risk-driver method:

The Quantik 0.7.1 Scenarios command runs these what-ifs together on the same random numbers (10,000 iterations, seed 42, ρ 0.3 between sections). In 0.7.1 the P10/P90 reading also applies to risk impacts:

The correlation test puts numbers on it: at P90 the gap between ρ = 0 and ρ = 1 is $6,817,065, and the line-item model with 0.3 between sections sits 42% of the way across it.

03.3 · The risk-driver result

With risk drivers (47 inputs, ρ 0.3 within and between sections, 10,000 iterations in 19 seconds):

03.4 · Schedule and joint confidence

With the Schedule linked, time-related items follow the simulated duration and programme-delay risks add weeks instead of being priced as weeks × rate.

03.5 · Top risks

Note the gap between expected value and worst case on P06 and P08: these are the lines to check against the LPHC screening thresholds.

04 · Authority templates

Folder: examples\authority templates. Each template is the full BOQ risk model set up the way one authority asks, with a summary sheet in that authority's format, worked on the highway job. Replace the BOQ, register, drivers and programme with your own project's.

Every template includes Risk Model, Risk Drivers and Schedule set to the authority's method; ρ 0.3 between sections; the correlation test; P90 precision; a run log; a P5–P95 table at 5% steps, unescalated; and the Results sheet with tornado, regression and rank table.

04.1 · Commonwealth: DITRDCA (GN3A)

Method: risk drivers (GN3A prefers risk factors). Schedule linked. Escalation outside contingency (GN4).

Also: base estimate by PCB Level 1–3 category reconciled to the BOQ, annual cash flow by financial year in base-date and outturn dollars, and a GN3A checklist.

04.2 · Queensland: TMR PCEM 9th edition

Method: line-item ranging ("planned" risk) plus the register ("unplanned"). Funded level P90 (QTRIP budget). Summary in the Annexure L / Form M4755 layout.

The estimate category check flags the 11.8% P90 contingency as below the expected range for a Category 3 (business case) estimate in PCEM Table 3.3.7: exactly the prompt a reviewer would raise. The sheet also carries contingency monitoring (§10.1.21) and the certification block.

04.3 · NSW: INSW Cost Control Framework

Method: risk drivers. Tier 1 (HPHR), funded at P90. Schedule QRA required and linked.

04.4 · Victoria: DTF RTCC

Method: risk drivers (HVHR business case: parametric with probabilistic risk events). Escalation sits inside the base risk estimate.

The TEI confidence level is assumed to be P90 because the RTCC guideline does not name it; confirm with DTF. The project / department / DTF split of the excess risk estimate uses placeholder percentages.