Why Good Decisions Often Look Wrong Under Uncertainty
Projects often assume the purpose of QCRA is to identify the best option.
But under uncertainty, that assumption can become surprisingly dangerous.
Because the option that produces the best number is not always the option that produces the best outcome.
The lowest expected cost.
The shortest expected duration.
The smallest contingency.
The highest expected value.
These can all appear attractive on paper.
Yet projects are rarely delivered under average conditions.
They are delivered in environments shaped by uncertainty, changing assumptions, and emerging constraints.
That distinction matters.
Earlier articles in this series explored how uncertainty can become distorted through:
distribution choices,
behavioural simplifications,
scenario structures,
governance translation,
and the disconnect between modelling outputs and real decisions.
Those articles largely focused on how uncertainty is represented, interpreted, and communicated.
This article shifts the focus toward a different question:
Even when uncertainty is represented well, what actually makes a good decision under uncertainty?
Because identifying exposure is only part of the challenge.
Projects must still decide what to do about it.
And in many cases, the decision that appears strongest in a model can be surprisingly fragile in reality.
Optimisation and Decision Quality Are Not the Same Thing
One of the most common assumptions in project decision-making is that the option producing the most attractive output is automatically the best option.
Often that means selecting the option that delivers:
the lowest expected cost,
the shortest expected duration,
the highest expected benefit,
or the most favourable percentile outcome.
At first glance, this seems entirely rational.
Projects should seek better outcomes.
The problem is that optimisation and resilience are not always aligned.
An option can appear highly attractive under expected conditions while becoming extremely vulnerable when assumptions begin to fail.
In many cases, the decision that performs best in the model is simply the decision that performs best if the future behaves as expected.
Unfortunately, uncertainty exists precisely because the future may not behave as expected.
Projects Are Delivered Once, Not Thousands of Times
One of the more subtle misunderstandings in quantitative risk analysis is the tendency to think in averages.
Monte Carlo simulation may run thousands of iterations.
The project experiences only one.
That distinction is important.
Expected values are useful.
Percentiles are useful.
Sensitivity analysis is useful.
But projects do not experience statistical averages.
They experience a single path through reality.
A strategy that performs marginally better under expected conditions may perform significantly worse when confronted by:
interface disruption,
procurement instability,
productivity deterioration,
access constraints,
stakeholder delay,
or cumulative delivery pressure.
The objective is not simply identifying what performs best most of the time.
It is understanding what remains viable when conditions become less favourable.
Robustness Versus Efficiency
Many project decisions involve a trade-off between efficiency and robustness.
Efficient strategies often seek to:
minimise cost,
minimise duration,
reduce contingency,
eliminate apparent redundancy,
and maximise utilisation.
Robust strategies often seek to:
preserve flexibility,
maintain recovery options,
absorb disruption,
reduce dependence on critical assumptions,
and protect delivery resilience.
These objectives are not always compatible.
A highly optimised plan may leave little room for recovery.
A highly efficient programme may depend on assumptions that prove fragile under pressure.
Conversely, a more robust strategy may appear less attractive because it carries:
additional cost,
additional time,
additional flexibility,
or additional contingency.
The irony is that these apparent inefficiencies are often what preserve delivery performance when uncertainty materialises.
Why Good Decisions Can Look Worse on Paper
This is where many organisations become uncomfortable.
Robust decisions frequently look inferior when evaluated using simplistic measures.
For example:
a packaging strategy may increase baseline cost while reducing downside exposure,
a procurement approach may appear slower while reducing execution risk,
a schedule allowance may increase duration while improving delivery confidence,
a larger contingency position may reduce affordability while improving resilience.
On paper, these decisions can appear less attractive.
In practice, they may significantly improve the probability of successful delivery.
The difficulty is that governance systems often reward optimisation more readily than resilience.
Lower numbers are easy to explain.
Protection against futures that may never occur is often harder to justify.
The Hidden Cost of Optimising for Approval
Projects rarely operate in neutral environments.
They operate under:
funding pressure,
affordability constraints,
governance expectations,
executive scrutiny,
and delivery commitments.
Under those conditions, optimisation can gradually become focused on approval rather than resilience.
Strategies that improve affordability often become more attractive than strategies that improve adaptability.
Assumptions become tighter.
Allowances become smaller.
Flexibility becomes harder to defend.
The project may become more attractive to approve while simultaneously becoming less resilient to uncertainty.
This does not occur because organisations are irrational.
It occurs because governance systems naturally reward confidence, affordability, and clarity.
The challenge is ensuring those incentives do not unintentionally undermine resilience.
Optionality Has Value
One of the most underappreciated concepts in project delivery is optionality.
Optionality is the ability to adapt when circumstances change.
Examples may include:
alternative delivery pathways,
packaging flexibility,
procurement options,
schedule float,
contingency reserves,
or additional recovery mechanisms.
These often appear inefficient when uncertainty does not materialise.
But uncertainty is precisely the reason they exist.
Projects frequently discover the value of optionality only after it has been removed.
Once flexibility disappears, recovery becomes significantly more difficult.
The Most Dangerous Decision
The most dangerous decision is often not the one with the highest exposure.
It is the one that appears highly attractive because everything must go according to plan.
The more a strategy depends on optimistic assumptions remaining true, the more vulnerable it becomes when conditions change.
This does not mean projects should avoid ambition.
Nor does it mean projects should always choose the most conservative option.
It simply means uncertainty should influence how decisions are evaluated.
Not only how outcomes are calculated.
What Good Looks Like
Good QCRA should do more than identify which option produces the best expected outcome.
It should help decision-makers understand:
which assumptions matter most,
where resilience begins to deteriorate,
which exposures dominate downside outcomes,
how alternative strategies behave under stress,
and what flexibility exists when conditions change.
The objective is not merely to optimise outcomes.
It is to improve decision quality under uncertainty.
That distinction matters.
Because projects do not succeed because they selected the most attractive number.
They succeed because they selected strategies capable of surviving reality.
Closing Thought
Projects often optimise for the future they hope will occur.
Resilient projects prepare for the futures that might occur.
The difference may appear small inside a model.
But it can become enormous during delivery.
Because under uncertainty, the best-looking decision is not always the best decision.
And sometimes the strongest strategy is the one that appears slightly less efficient, but remains viable when assumptions begin to fail.
