Scientists have discovered why complex problems are easier to solve in parts

The study was published in Nature Human Behaviour. In it, participants played a simple weather forecasting game: they had to guess whether it would be "sunny" or "raining" using geometric shapes. Some of them were first trained on individual figures, while others were shown combinations of three figures at once. In the end, the first group did better, even when everyone was given the same difficult task later.

Details

In the experiment, participants were shown simple shapes such as circles, triangles and squares. Each figure was associated with a certain probability of response: some were more likely to suggest sunshine, while others were more likely to suggest rain.

The task looked simple, but it had an important complexity. In a real test, a person had to take into account not just one figure, but several at once. That is, the participant had to understand how the different cues worked together.

The researchers compared the two ways of learning. In the first case, people first saw only one figure at a time and gradually memorised what it meant. In the second case, participants were immediately shown a set of three figures - that is, a more complex whole picture.

The first option won. People who started with separate clues coped better with the final tasks, where the figures appeared together. In other words, learning "piece by piece" did not prevent them from seeing the big picture, but on the contrary helped them to assemble it faster.

To explain the result, the scientists built a computer model. It showed that a person probably uses not just one strategy, but a combination of two. Sometimes he evaluates each feature separately, and sometimes he combines several features into one common conclusion.

Simply put, if a task is too complex, it is useful to remove the unnecessary noise first. Deal with one element, then a second, then a third - and only then put it all together. It's like a jigsaw puzzle: it's harder to start with a chaotic pile of pieces than it is to find clear pieces first.

The researchers also tested whether the order of learning has an effect. It turned out that it helps people to start with more obvious examples. When simple and straightforward cases are given first, and then more subtle and ambiguous cases are gradually added, learning goes faster.

Why it matters

The conclusion sounds simple, but it's important for learning. Complex topics often fail not because a person is 'unable to understand', but because they are given too much information at once.

This doesn't just apply to school or university. This is the way to learn a new language, programming, driving, working with data, medical diagnosis, playing an instrument, or any skill where you need to consider several rules at once.

Research does not mean that any material should always be crushed to infinity. In real life, individual elements often depend on each other, and you still need to learn how to connect them. But the work shows: a good start is not an overload, but a clear sequence from simple to complex.

Background

The idea of "first simple, then complex" has long been used in education. This is how textbooks, training and professional courses are structured. But it's not always clear why this approach really works and in which tasks it provides an advantage.

A new study adds experimental evidence to this idea. Researchers didn't just ask people how they were more comfortable learning, they compared two learning scenarios in the same task.

The result supports an approach often referred to as learning by programme or by stages: first a person learns the basic elements, then encounters more complex combinations. This order can reduce overload and help one understand a rule more quickly.

Source

Qingtian Mi et al, "Human curriculum learning of a cue combination task", Nature Human Behaviour, 2026.

In the study, participants took online experiments with a cue combination learning task. They had to predict an outcome based on geometric shapes. The scientists compared learning from individual cues and learning from combinations and then tested which strategy was better at solving complex tasks.