TL;DR: We introduce the "Decisions Over Time (DOT) task" to measure active inference, the cognitive skill of updating beliefs in uncertain situations. The task unexpectedly revealed two very different successful strategies for solving complex problems: "Adapters," who relentlessly seek perfect, underlying rules, and "Satisficers," who find and execute efficient, "good-enough" shortcuts.

The 'Orient' Problem: Why We Get Stuck

Modern leaders operate in environments defined by volatility, uncertainty, complexity, and ambiguity (VUCA). In these situations, past experience can be a dangerous trap: it makes you cling to outdated models of the world.

For decades, military doctrine has used models like Colonel John Boyd's OODA loop (Observe, Orient, Decide, Act) to describe the ideal decision-making process. The problem? The "Orient" phase (the part where you actually make sense of the world to form a mental model) is the most critical, yet the most misunderstood and hardest to train.

How do you train someone to "orient" when they land in a situation they've never seen before? You can't give someone a checklist for disorientation. The only training target is the learning process itself.

Active Inference: The Brain's Strategy for Minimizing Surprise

Our research proposes that this missing competency is active inference.

Active inference is a theory from cognitive neuroscience that describes a core mechanism of the brain. It posits that your brain is not a passive receiver of information but an active, predictive machine. Its primary goal is to minimize "surprise" (or prediction error), which is the mismatch between what you expect to happen and what actually happens.

It's an iterative process running constantly: you predict what happens next, act, observe the outcome, and if it surprises you, update your model. That update is the signal to learn.

This is exactly how we make sense of a new, fast-changing situation.

Measuring the Unmeasurable: The Decisions Over Time (DOT) Task

To measure this, we developed a new computerized test: the Decisions Over Time (DOT) task (shown above).

Imagine a simple game. On each trial, you see three colored dots (red, green, and blue) in a row. Your only instruction is to "click the correct dot". The "correct" dot is determined by a hidden rule based on the color of the middle dot. For example, "Rule 1" might be: If the middle dot is Blue, the Left dot is correct; if Red, the Right is correct; if Green, the Middle is correct.

You only learn from simple feedback: a "correct" click gives you a point, an "incorrect" click loses one. You must use this feedback to infer the hidden rule.

But here’s the twist. Just when you master the rule and start scoring perfectly, the task changes the rule without telling you. Suddenly, your perfect mental model is wrong. You start getting "surprising" feedback (errors). The task forces you to inhibit your old, outdated belief and actively infer the new rule.

The Surprising Finding: Two Paths to Success

We didn't just find "good" and "bad" performers. We found four distinct groups, including two very different types of successful performers.

1. The Adapter: The "Perfect Solution" Strategist

This is the "ideal" performer we expected. Adapters are logical and relentless. They used the feedback to learn Rule 1. When the rule changed, they detected the new errors, discarded their old model, and learned Rule 2, then Rule 3, and so on. This is a high-effort, high-focus, executive functioning approach. They work hard to eliminate error and find the single "correct" solution.

2. The Satisficer: The "Good-Enough" Strategist

This finding was a complete surprise. Satisficers never learned the hidden, complex rule at all.

Instead, they quickly identified a "good-enough" statistical shortcut. They noticed that, due to the task's design, one color happened to be the correct answer 66.6% of the time, regardless of its position. So, they just clicked that one color on every single trial.

They willingly accepted a 33% error rate because this simple, low-effort strategy was fast, efficient, and still resulted in a high score. They adopted a probabilistic learning approach, finding a solution that was correct two-thirds of the time and could be executed 100% of the time.

Why This Matters: From "Perfect" Plans to "Good-Enough" Now

The Adapter vs. Satisficer split shows up in real decisions too.

It brings to life General George S. Patton's famous quote: "A good plan violently executed Now, is better than a perfect plan next week."

• Adapters are the "perfect plan" strategists. Their deep, analytical approach is essential for high-stakes, complex problems where "good-enough" is catastrophic and getting the answer right is all that matters.
• Satisficers are the "good plan now" strategists. Their fast, efficient, probabilistic approach is vital for dynamic, fast-moving situations where speed and resource management are more important than perfection.

The problem is that leaders often have a default style, but the world demands both. An effective leader must know when to be an Adapter and when to be a Satisficer.

The DOT task could work as a diagnostic tool that reveals a leader's natural tendency. Once leaders know their own style, they can learn to switch strategies deliberately and match their approach to the problem in front of them. It gives us a concrete, measurable way to develop the "Orient" competency, something doctrine has described for decades without ever operationalizing.

Full Citation

Nye, J. M., Stothart, C., Graves, R., Francisco, A., & Peterson, J. (2024). Active inference: A competency for making decisions in uncertain situations (ARI-SS-SR 2024-09). U.S. Army Research Institute for the Behavioral and Social Sciences.