RES Contingency Guideline (3rd Ed 2025) Explained

National Best-Practice Authority

What the RES Guideline Is — and Why It Carries Weight

The Risk Engineering Society is a Technical Society of Engineers Australia, the country's largest professional engineering body. That gives the guideline cross-jurisdiction authority no single state framework holds.

The RES Contingency Guideline is the de-facto Australian reference for cost and time contingency on capital projects. Because the RES sits within Engineers Australia, it speaks for the profession nationally rather than for one state agency. The 3rd Edition (2025) is a 224-page document authored by RES National President Pedram Danesh-Mand and is endorsed by the UK Institute of Risk Management through its Infrastructure Risk Special Interest Group.

Critically, it is a general guide, not a mandatory standard. The guideline explicitly states that it “does not set mandatory or minimum standards.” Instead, it provides the method best practice that binding jurisdictional mandates — such as Queensland's TMR risk and contingency process and the Commonwealth (DITRDCA) cost-estimation guidance — assume but do not themselves teach. It is the “how” sitting beneath each jurisdiction's “what and when.”

Edition history (use these dates)

1st Edition — May 2016
2nd Edition — February/March 2019
3rd Edition — 2025 (the current edition)

What's new in the 3rd Edition: the RES / Western Sydney University Contingency Management Professional (CMP) certificate and masterclass; expanded Expected Accuracy Range and Base/Best/Central Estimate definitions; updated software and AI-based tools; enhanced Quantitative Schedule Risk Analysis (QSRA); inclement-weather risk guidance; the new Contingency X-Factor concept across development and delivery phases; pre-tender-estimate accuracy factors; and major new appendices on Australian federal/state contingency requirements (Appendix G) and international practice (Appendix H).

The Heart of the Guideline

Base, Contingency, Management Reserve & Escalation

RES defines contingency as a “specific allocation of resources required in addition to the Base Estimate or Base Schedule as a provision for inherent and/or contingent risks for the desired confidence level.” The discipline lies in keeping four things separate.

Component	What it covers	Where it sits
Base Estimate / Base Schedule	The cost/time reasonably expected if the project proceeds as planned. Excludes escalation, foreign exchange, contingency and management reserve.	The foundation
Contingency	Covers inherent and contingent risk within the defined scope. Expected to be expended. Typically the P50 portion once allocated.	Inside the PMB
Management Reserve (MR)	The gap above the funded confidence level (e.g. the P90−P50 portion) and unforeseen in-scope work. Released by higher delegation.	Outside the PMB
Escalation	Anticipated cost growth over time (inflation, market). RES discusses its relationship to contingency but excludes its determination from scope.	Separate — out of RES scope

Contingency is expected to be spent

It funds inherent and contingent risk within the defined scope. It is not a discretionary buffer and never a substitute for proper estimating or planning.

Management Reserve answers to a higher delegation

Held outside the PMB for the gap above the funded confidence level and for unforeseen in-scope work — not controlled by the project manager.

What contingency is NOT (§5.2)

It expressly excludes major scope changes, acts of God (earthquakes, floods), project funding delays, and changes to delivery strategy. These are handled commercially, by insurance, or by re-baselining — never buried in contingency.

Deterministic → Probabilistic → Simulation

The Methods, and How RES Ranks Them

RES structures methods for both cost and schedule — and insists the two be analysed in an integrated way (§5.3) — from simple deterministic percentages up to the recommended simulation tier. There is “no one method to suit all situations.”

The explicit ranking (§5.4.1)

RES recommends bottom-up probabilistic methods “wherever possible at every decision gate … particularly at the full funds approval gate.” For projects valued over $10M, the guideline recommends simulation or hybrid methods and “generally recommends the application of MCS [Monte Carlo Simulation].” Method choice depends on lifecycle stage, project value and risk, scope-definition maturity, complexity and organisational maturity.

Deterministic (early-stage / small projects only): pre-defined factor-based %, expert-judgment %, item-based %. RES gives example tables but “does not recommend these, or any other specific values.”
Probabilistic, non-simulation (Appendix D): Enhanced Scenario-Based Method (eSBM), Method of Moments (not recommended for construction/infrastructure), Reference Class Forecasting, Parametric, Regression, Range-based, AI-based.
Probabilistic, simulation (the recommended tier): Quantitative Cost Risk Analysis (QCRA — including three-point, risk-factor and First Principles Risk Analysis / FPRA), Quantitative Schedule Risk Analysis (QSRA), and integrated Quantitative Schedule-Cost Risk Analysis (iQSCRA).

The Technical Detail

Inside a Probabilistic Model

When RES recommends simulation, it is specific about how to do it well: deliberate distribution choice, honest ranges, and — above all — correlation.

Three-point estimates and distributions (Appendix E)

Three-point estimates (min / most-likely / max per WBS element) feed QCRA. RES warns the bare three-point method “may result in inconsistent and unpredictable results” and is not recommended at key decision points (Final Business Case / FID) — RES prefers FPRA there. For the distributions themselves, RES lists continuous forms (Triangular, PERT/BetaPert, Uniform, Normal, Lognormal, Beta) and discrete forms, and recommends the “Alt” variants — notably AltPert over PERT — because subject-matter experts rarely capture the true worst and best cases. Absent objective data, RES recommends modelling SME bounds with 80% confidence intervals (10% / 90%), adjusted for skew.

Correlation matters more than the distribution choice (§12.7)

RES distinguishes functional/implicit correlation (mathematical relationships in the model) from applied/explicit correlation (analyst-specified coefficients from −1 to +1). Its strongest modelling warning: ignoring correlation can understate total-level variance “considerably,” and “the effect of excluding correlations … is more profound than the choice between different probability distributions.”

The five risk sources to address (§5.2)

Inherent risk in the estimating/scheduling method itself.
Inherent technical/programmatic (systemic) risk — including engineering and design-maturity uncertainty, which is always present.
Contingent risks at interfaces.
Correlation risk.
Specific project risks.

Cost and schedule are interlinked, not separable. iQSCRA cost/resource-loads the base schedule so that time risk flows into cost. The Merge Bias Effect — parallel logic paths converging at a milestone — explains why deterministic schedules tend to run optimistic.

P-values & Confidence Levels

The Average Is Not the 50/50 Point

P(x) is the estimate with an x% probability that cost or time will not be exceeded. The most useful insight RES offers clients: the mean of a project cost distribution typically sits well below the median.

P50 with the Base / PMB

RES's working convention pairs P50 with the baseline and P90 with the Management Reserve at a higher delegation.

Report the S-curve at P10, P50 and P90

FPRA reports must present the S-curve at all three percentiles and allocate contingency at both P50 and P90.

The funded percentile is a risk-appetite decision

RES does not mandate a single funded percentile — it is framed as a function of organisational risk appetite, tolerance and capacity, reassessed at every decision gate (and, for government work, the relevant jurisdiction's requirements).

The Outside View

Reference Class Forecasting — Done Right

RES covers reference class forecasting (RCF) in Appendix D — attributed to Kahneman & Tversky, Lovallo & Kahneman, and Flyvbjerg & COWI (2004), and also called “optimism bias uplifts.” Its purpose is to counter optimism bias and strategic misrepresentation by basing an uplift on the cost-overrun distribution of similar past projects.

RES's key positioning

RCF is “more properly considered a validation or benchmarking practice for quality management and governance, NOT a contingency determination method,” and RES “recommends this method to be used to complement rather than replace … First Principles Risk Analysis (FPRA) supported by QSRA.”

RES lists seven limitations: it assumes the future mirrors the past; discourages improvement; most organisations lack good data; needs a genuinely similar reference class; ignores project-specific contingent events; and ignores safety, quality, environment and reputation. It reproduces the UK Supplementary Green Book optimism-bias uplift table as a fallback only.

The right way to use it: a governance and benchmark cross-check on bottom-up Monte Carlo results — never a substitute for them.

Appendix G — The Connective Tissue

Three Frameworks, One Method

The 3rd Edition's Appendix G (“Australian Government and Contingency”) compiles each jurisdiction's own cost-estimation and contingency requirements. RES is explicit it is not setting the rules — it is the method layer beneath the binding mandates, with the stated aim of helping “reach consensus in the methods used for contingency determination across industry and governments.”

The value of this single appendix is that RES has already done the cross-jurisdiction mapping — and it reveals that every Australian jurisdiction is converging on the same probabilistic, P50/P90 method RES describes. Cenex delivers that method to whichever jurisdiction's mandate applies.

Jurisdiction	Framework	Method & confidence level (per Appendix G)
Commonwealth — Dept of Finance	Two Stage Capital Works Approval	P50 at First Stage, P80 at Second Stage; Monte Carlo S-curve required.
Commonwealth — DITRDCA	Infrastructure Investment Program	Probabilistic mandatory for total out-turn (incl. contingency) > $25M; P50/P90; PCB templates. See our DITRDCA guide.
Queensland — TMR	PCEM + Risk & Contingency Process Manual	P50 and P90 in the PCB; probabilistic mandatory for federally funded projects with out-turn P90 > $25M. See our TMR guide.
New South Wales	Infrastructure NSW Cost Control Framework	Contingency identification mandatory for projects > $100M; Monte Carlo minimum for High-Profile/High-Risk; TfNSW requires P50 and P90; generic % “to be avoided.”
Victoria	DTF Resource Management / HVHR + OPV RTCC	Approval at the P90 estimate; contract award reset to the P60 estimate; central-contingency release model.
Western Australia	Treasury SAMF	Probabilistic (P90 or P50) for high-value/high-risk; buildings guideline uses deterministic layered contingencies (Planning 5% / Design 10% / Construction 10%).
South Australia	Treasury TI17 + DIT EST600 Estimating Manual	Full compliance ≥ $11M (±10% assurance); probabilistic P50 and P90 required.
Tasmania	SIIRP	Contingency documented in the business case; no prescribed confidence level or threshold.
Northern Territory	—	RES found no specific contingency-management reference.
ACT	Capital Framework (tiered)	Tier 1 must use stochastic (Monte Carlo) techniques; Tier 3 deterministic acceptable.

Appendix H extends the picture internationally — notably the US FTA/FRA Oversight Procedure 40 design-stage minimum-contingency drawdown ladder, NZ Gateway QRA, and the UK IPA's Reference Class Forecasting plus QRA. The consistent thread is the same one RES advances at home: probabilistic, bottom-up, benchmark-validated contingency. For a side-by-side of the three Australian frameworks Cenex works under most, see Frameworks Compared.

Governance, Competency & Bias

Bias Control & Independent Review

RES anchors contingency governance in board-level risk appetite (citing AICD guidance): “Uncertainty does not justify a lack of discipline or integrity.” The estimate and schedule must be validated and bias-assessed before contingency is finalised.

Thirteen biases, tabulated

RES Table 2 names anchoring, availability, groupthink, blind-spot, confirmation, overconfidence, optimism, recency, commitment, framing, sunk-cost, hindsight and planning-fallacy biases — each with a mitigation.

Independent third-party review

A headline mitigation: benchmarking, anonymous interviews, wide upside-biased ranges, P-ranges over single points, and — critically — an independent reviewer with no downstream delivery interest.

Many short reviews, not one long one

RES recommends cross-checking assumptions across separate meetings rather than a single long workshop — a deliberate guard against groupthink and anchoring.

Recognised competency

RES recognises the RES/WSU Contingency Management Professional (CMP) certificate and Chartered Professional Engineer (Risk or Cost Engineering), plus AACE (CCP, CEP, DRMP), PMI, IRM and ISO 31000 credentials.

How Cenex Delivers It

RES Best Practice, Applied

Cenex's risk practice is built around the RES recommended tier — bottom-up First Principles Risk Analysis with Monte Carlo simulation — delivered to whichever jurisdiction's mandate funds the project.

First-principles + probabilistic. Base estimate built bottom-up; risks modelled in @RISK with 10,000 iterations; correlation modelled deliberately; distributions chosen on purpose (AltPert / Triangular); S-curve reported at P10/P50/P90.
RCF used as RES intends. Reference class forecasting is a benchmark and governance cross-check on the modelled result, not a shortcut.
Active bias management. Optimism, anchoring and planning-fallacy biases countered through independent review, separate-session assumption testing, and P-range inputs — the discipline that distinguishes credible contingency from a number pulled to fit a budget.
Chartered sign-off. Cenex's Chartered-Engineer review matches the competency RES recognises for contingency work.

Explore the rest of the hub: TMR Risk & Contingency, DITRDCA Cost Estimation, Frameworks Compared, or our Risk Modelling & Management service.

Need RES-Aligned Contingency You Can Defend?

Cenex delivers independent, first-principles probabilistic contingency — Monte Carlo, P50/P90, correlation modelled, benchmark-validated — aligned to RES best practice and your funding jurisdiction's mandate.

Talk to Cenex Our Risk Service

The RES Contingency Guideline (3rd Edition, 2025)