COMMUNITY BRIEFING NOTE 09
Preference to “fact” – How robust is the Matilda bay Multi Criteria Analysis?
16 September 2026 — Damien Smith, Naval Architect
If the Government says the answer is clear, can we see the work?
The Government has repeatedly relied on its multi-criteria analysis — the MCA — to justify choosing Central Matilda Bay for the new UWA ferry terminal.
Transport Minister Rita Saffioti has gone further. In Parliament, while discussing the MCA, she has repeatedly described the Government’s case in terms of “facts”. In February she said:
“Facts are important in this discussion. Facts are very important.”
She then went through what she described as the facts assessing each option. [i]
At this point, it is worth taking a closer look at what an MCA actually is, and why Australian and international appraisal guidance is so cautious about treating the result of one as a fact.
An MCA takes a number of different considerations, assigns scores and weightings to them, and combines those judgements into a ranking.
That can be useful, particularly where some things are difficult to put into dollars. But the answer depends heavily on what goes into it: which criteria are chosen, how they are scored, how much weight they are given and whose judgement is used.
There is therefore plenty of room for what you would like to be true to find its way into the arithmetic.
Infrastructure Australia puts the problem more politely. It warns that MCA can be subject to “unconscious or advocacy bias”, and that scoring can be open to bias where people contributing to the analysis already prefer a particular intervention. It also warns against scores creating a misleading impression of accuracy or rigour. [ii]
That is why these methods come with rules about showing the evidence and testing what happens when the assumptions change.
HOW THIS WORK IS SUPPOSED TO BE DONE
WA Treasury, Infrastructure Australia and the UK Treasury Green Book approach the problem slightly differently, but the message is remarkably consistent.
WA Treasury itself labels MCA a “Possible Shortlisting Tool”. It sits in the Strategic Options Analysis and is one way of taking a long list of strategic options and filtering it down to the options which will receive further analysis. Treasury says the shortlisting process should be structured, objective and evidence-based, and that the method and criteria should be specified.[iii]
Infrastructure Australia similarly treats MCA as a filtering tool before more detailed analysis. It requires the objectives, criteria and weights to be documented, scores to be linked to evidence, and the results to be tested against changing assumptions. It also warns about double counting where the same impact influences more than one criterion. iii
The UK Green Book goes further. It now recommends against simple MCA — basic weighting and scoring — for decision-making because it can lack an objective basis and reduce transparency. Where a more rigorous MCDA is used, it is a longlist tool for complex technical trade-offs, not a substitute for detailed economic analysis.[iv]
Under that approach, criteria are developed and validated with experts and stakeholders. Genuine disagreement is recorded rather than buried in the score. Sensitivity analysis tests what happens when scores or weights change and whether the preferred option still comes first.[v]
There is also a useful local precedent. As discussed in Community Briefing Note 03, the 2004 analysis of what became the Mandurah railway discussed the problems with subjective weighting and double counting. Its economic evaluation instead used conventional quantified cost-benefit analysis. [vi]
So the broad best-practice picture is fairly clear.
MCA can help sort options. It is not a magic machine that turns judgement or preference into fact.
WHAT HAVE WE ACTUALLY BEEN SHOWN?
What Government has published is essentially a colour-coded table.
Interestingly, it looks much like the output side of the MCA spreadsheet template Infrastructure Australia itself provides — except the Government version does not contain the numbers.
Infrastructure Australia’s template provides for the criteria, weights, numerical option scores, scenarios, results and graphs.[vii]
The Government’s public version has the criteria and the colours: It does not have the numbers.
We can see that Central Matilda Bay gets good colours and JoJo’s gets bad colours. We cannot see the numerical scores behind those colours. We cannot see the weights or how they were chosen. We cannot reproduce the result or change an assumption and run it again.
In practical terms, we cannot tell whether JoJo’s finished narrowly behind Matilda Bay or nowhere near it.
Infrastructure Australia specifically says it is not sufficient simply to report the combined weighted scores and criteria scores. Enough information should be available to understand the evidence and judgements behind the result and whether the ordering survives sensitivity testing.
Here we do not even have the weighted scores.
If there is a full MCA sitting behind the attractive red, amber and green table, Government should publish it.
WHEN DID THE MCA APPEAR?
The timing is also worth looking at.
WA Treasury is quite clear about where an MCA belongs in the process.
Its Business Case Guidelines put MCA in the Strategic Options Analysis, under the heading “Multi-Criteria Analysis – Possible Shortlisting Tool”. Its purpose is to take a long list of strategic options and filter them down to the options which then receive the detailed evaluation needed to select a preferred option.
In other words, MCA is supposed to be near the front of the options process, not something added after the project direction has hardened.
The ferry project did not begin with five equal terminal locations and an MCA which then discovered Matilda Bay.
The Government’s own account says the original UWA masterplan placed the terminal in northern Matilda Bay. Additional locations were assessed “in response to feedback from user groups”.
So the five-location MCA appears to have arrived after Matilda Bay was already the working project direction and after stakeholders had started objecting.
NEW INFORMATION SHOULD GO BACK IN
The information available to Government did not stop changing when the MCA was completed.
The Waterways Stakeholder Reference Group was subsequently established to work through operational safety issues, and PTA says the final route will be determined with assistance from that group.
Paul Chivers’ subsequent independent comparative safety review has challenged important assumptions which favour Matilda Bay, particularly around whether the conflicts with existing recreational users can readily be managed.[viii]
The proposed operation has also continued to change. PTA has publicly said its service modelling was based on current gazetted speed limits while leaving open the possibility of exemptions. The operation now proposed involves speed exemptions being sought to achieve the preferred timetable, assumptions which are not visible in the published MCA.
These are exactly the sorts of changed assumptions that sensitivity analysis is for.
If JoJo’s waterways-safety assessment changes, does Matilda Bay still win?
If the effect on existing Matilda Bay users is greater than originally assumed, does Matilda Bay still win?
If the operating speeds and controls change, does Matilda Bay still win?
There is a technical name for the point where the answer changes: a switching value.
In plain English: How much does an assumption have to change before Matilda Bay stops winning?
WHAT WAS BEING COMPARED?
There is a bigger issue underneath all of this.
The terminal is not an independent decision.
By the time of the vessel tender, the system had already been framed around battery-electric ferries operating via Matilda Bay, high-power charging at Matilda Bay and vessels capable of operating at up to 25 knots.
Those choices affect the answer.
Change the vessel and you can change the charger. Change the charger and you can change the terminal footprint and cost. Change the speed and you can change the timetable, energy consumption, wake and safety interaction. Change the terminal and you can change the route, patronage and environmental impact.
The UK Green Book’s preferred longlist method deals with exactly this problem. Its Options Framework Filter effectively says: pull the proposal apart. Break it into component choices and recombine them in different ways to test genuinely different options.
For this project that means asking not only: Which terminal is best for this ferry system?
but: Which ferry system is best?
Did Government compare genuinely different combinations of terminal, vessel, speed, route, charging architecture and service pattern?
Or were most of those choices already fixed before the MCA was used to choose the terminal?
QUESTIONS FOR GOVERNMENT
1. Will the Government publish the underlying multi-criteria analysis used to select Central Matilda Bay, including the numerical scores, criteria weightings, basis for those weightings, stakeholder validation, sensitivity analysis and any switching-value or scenario testing undertaken?
If not, why not?
2. Has the MCA been rerun to incorporate information obtained since the original assessment, including Paul Chivers’ comparative safety analysis, the later views raised through the Waterways Stakeholder Reference Group, and the operation now proposed by PTA — including the speed exemptions now being sought and any resulting changes to vessel speeds, routes or operating controls?
If so, did any of those changes materially affect the scoring, ranking or selection of Central Matilda Bay?
If not, why is Government continuing to rely on the original MCA as evidence that Central Matilda Bay is clearly the preferred option?
About the author
Prepared by Damien Smith, Perth naval architect and Director of Damien Smith Design Pty Ltd.
Disclosures: Damien Smith Design was a tenderer for the Swan River ferry project. Damien is also on the Mounts Bay Sailing Club Management Committee and is an Australian Labor Party member. The views expressed are his own.
REFERENCES
[i] Western Australian Parliament — Legislative Assembly, 19 February 2026, Ferry terminal—Matilda Bay, Nedlands
[ii] Infrastructure Australia — Guide to multi-criteria analysis
[iii] WA Treasury — Strategic Asset Management Framework: Business Case Guidelines
[iv] HM Treasury — The Green Book 2026
[v] HM Treasury — Use of Multi-Criteria Decision Analysis in options appraisal of economic cases
[vi] Planning and Transport Research Centre — Perth’s South West Metropolitan Railway: Balancing Benefits and Costs, 2004
[vii] Infrastructure Australia — Guide to multi-criteria analysis and downloadable MCA template
[viii] Paul Chivers, Independent Safety and Risk Review — Proposed METRONET on Swan River Ferry Terminal Locations: Matilda Bay and Nedlands Jetty (JoJo’s), July 2026