Wicked problem

A wicked problem is a problem that is difficult or impossible to solve because of incomplete, contradictory, and changing requirements that are often difficult to recognize. The term was coined by Horst Rittel and Melvin Webber in 1973 to distinguish problems that are tame — definable, solvable, and stable — from problems that are wicked — ill-defined, unsolvable in any final sense, and constantly evolving.

Wicked problems have no stopping rule. You do not solve a wicked problem; you merely reach a point where you decide to stop working on it. They have no test of success: there is no way to know whether a solution is correct, because the problem itself is not fully defined. Every attempt to solve a wicked problem is also an attempt to define it, and every definition reveals new aspects that were not visible before. The problem and its solution co-evolve.

The classic examples of wicked problems include urban planning, climate change, poverty, and healthcare reform. But the concept applies equally to technical systems. The design of a complex adaptive system is a wicked problem: the system's requirements change as it is being designed, because the users adapt to the system and the system adapts to the users. The optimization of a financial network is a wicked problem: improving the resilience of the network may reduce its efficiency, and reducing its efficiency may make it less resilient in unexpected ways. The CDS market is a wicked problem: the optimization of individual risk transfer produces systemic fragility, and the elimination of systemic fragility requires suboptimizing individual risk transfer.

Wicked problems are characterized by emergence: the system acquires properties that were not designed and cannot be predicted. They are characterized by path dependence: the history of attempted solutions constrains the space of future solutions. They are characterized by irreversibility: every intervention changes the system in ways that cannot be undone. And they are characterized by pluralism: there is no single correct perspective on the problem, because different stakeholders define the problem differently.

The standard tools of engineering — requirements analysis, optimization, verification — fail when applied to wicked problems. The appropriate tools are not algorithmic but adaptive: iterative design, stakeholder negotiation, scenario planning, and the acceptance of uncertainty. The goal is not to solve the problem but to manage it: to steer the system toward desirable states while accepting that the steering itself changes the system.

The concept of wicked problems is a corrective to the hubris of rationalist problem-solving. It says: some problems are not waiting to be solved; they are waiting to be lived with. The CDS market, the climate, the city — these are not puzzles with correct answers. They are processes with tradeoffs, and the tradeoffs are permanent.