OODA Loop: Decision-Making Under Bounded Rationality

The decision never occurs under ideal conditions.

In my article “The Decision Maker and External Memory: The new equation of competitive advantage: Extended Rationality”, when approaching the topic of limited rationality, I wrote that, according to Herbert Simon, complex circumstances, limited time and inadequate mental computational power reduce the decision maker to a state of “limited rationality” ^[1].

This formulation continues to be one of the most relevant for understanding the contemporary decision-making process.

Simon argues that, faced with complex situations – often beyond their full understanding – the decision-maker uses their information processing capabilities not to find the best possible solution, but to find a solution good enough for the context in which they find themselves. ^[2]. The decision-maker does not optimise; they satisfice. They seek alternatives, estimate consequences, reduce uncertainty and, where possible, identify a course of action that is operationally acceptable.

The implication is structural.

The decision maker does not know all the available alternatives or all their consequences. You only follow one path and you will never know if that path is the best. At most, you can make a reasonable guess under certain conditions ^[1]. The limitation is not just cognitive; It's material. Time and cost prevent the collection of all relevant information. The decision-making process is inevitably incomplete ^[2].

What the decision maker does, in practice, is build expectations about the future based on known empirical relationships and the information available at the time. These expectations are not certainties. They are approximations.

It is in this context that the Decision Cycle, or OODA Loop – Observation, Orientation, Decision and Action – developed by John Boyd, colonel in the United States Air Force, emerges. ^[3].

The OODA Loop does not solve bounded rationality.
But it makes it operable.

Boyd arrived at this formulation from empirical observation during the Korean War. The United States used F-86 Sabers against the Soviet MIG-15s, which were, in several technical aspects, superior – faster, more powerful and with better climbing capacity. Still, American pilots achieved a kill ratio of approximately 10:1.

The advantage was not in the machine.
I was in the process.

Boyd identified two critical factors: greater pilot visibility – that is, better situational awareness – and an aircraft that is easier to maneuver, less physically demanding. These elements allowed American pilots to observe, interpret and act more quickly than their adversaries.

From this observation came one of the most important ideas in decision theory: the advantage is not in having more information or better resources, but in completing decision cycles more quickly than the opponent.

The OODA Loop formalizes this process.

Observing involves collecting relevant information and building a representation of the environment. But observing is not accumulating data. It is to distinguish between signal and noise. Just as in an emergency medical context, where the clinician collects signals, interprets patterns and discards irrelevant information, the quality of observation depends on the ability to select what matters.

Orienting is the central phase.

Boyd described it as schwerpunkt – the point of greatest effort. This is where information is interpreted. And this is where limited rationality manifests itself with greater intensity. Guidance depends on mental models, experience, culture and analytical skills. Two decision makers can look at the same data and come to radically different conclusions.

Boyd identified four main barriers to mentoring: cultural tradition, genetic inheritance, capacity for analysis and synthesis, and the continuous flow of new information. These barriers are no exceptions. They are permanent conditions of the decision-making process.

Guiding yourself correctly does not mean eliminating bias.
It means operating in spite of it.

The decision comes only after this process. But it is not an end point. It's an intermediate point. The decision should be understood as a hypothesis – not as a conclusion. And, therefore, it must be open to continuous review.

This is where a Bayesian logic implicitly comes in: each decision must be tested against reality and adjusted based on new information.

Acting, finally, is the moment when the decision materializes. Without action, there is no feedback. And without feedback, there is no learning.

This is precisely why the concept of “loop” is essential.

The OODA Loop is not a linear process. It's a continuous cycle. Each action changes the environment, generates new information and restarts the process. The advantage is not in completing a perfect cycle, but in completing successive cycles faster than others.

In an environment of limited rationality, this speed becomes decisive.

The decision cannot wait for perfect information.
It has to occur within the imperfection.

This principle has profound implications outside the military context. In organisations, markets or individual decisions, the dominant temptation is to prolong the analysis phase. Look for more data, more validation, more certainty.

But certainty is a luxury that rarely exists.

The OODA Loop proposes an alternative: decide based on the best information available, act, learn and adjust. Learning does not precede action. It results from it.

As Boyd himself stated, we cannot rely solely on our past experiences nor apply the same mental models indefinitely. It is necessary to cross disciplines, integrate new perspectives and continually adapt the decision-making process to the context.

The advantage is not in avoiding error.
It's about learning faster than others.

In a world characterised by complexity, speed and imperfect information, the decision maker is not the one who knows more, but the one who can iterate better.

And it is in this ability to iterate – within the limits of rationality – that the true advantage lies.

Editorial note

This article was originally published by Rui Silva in MIT Technology Review Portugal and has been adapted to the Golden Blue Notes editorial style, fully preserving the logic, examples and original content.

Golden Blue Notes
2BePro · Central Brain Trust · PIM Trifecta