Skip to main content
London Tutorsby KnowledgeDottCom

19 November 2026 · London Tutors

When Maths Cannot Predict the Future — Until It Can

Illinois physicists built a simple grid of cells that follows fixed rules, yet no computer could predict how it would end up — until, as it ran, a pattern appeared that made the ending almost perfectly predictable.

When Maths Cannot Predict the Future — Until It Can

If a system follows fixed rules with no randomness at all, you might expect its future to be easy to work out. Physicists at the University of Illinois have shown that is not always true — and then found something even stranger: a system that starts out impossible to predict can, as it runs, grow the very pattern that makes its ending predictable.

The work, published in the journal Nature Communications on 11 September 2026 and led by Professor Hyun Youk with graduate student Elinor Kay and lead author Lars Koopmans, uses a "cellular automaton" — a grid of cells, each in one of a few states, that changes step by step according to simple fixed rules. The team's grid was inspired by living tissue, where cells send out chemical signals and sense the signals of their neighbours.

The researchers set machine-learning algorithms a simple job: look at the starting grid and predict which of three endings it would reach — a frozen, static pattern, a wave moving in a straight line, or a moving spiral wave. The algorithms did no better than guessing. Half the time they were right, half the time completely wrong — no better than tossing a coin. The system was fully deterministic, with no chaos and no randomness, yet genuinely unpredictable from its starting point.

Then the team found the twist. As the grid ran, its cells organised themselves into a hidden structure — a "winding field" that captures how the pattern twists and turns. When the researchers fed that winding field to the same machine-learning system, its accuracy climbed as the simulation went on, from no better than guessing at the start to almost perfect by the end. Predictability, in other words, was not there at the beginning — it was built by the system itself as it ran.

Why does this matter? Most people meet unpredictability through chaos — the famous "butterfly effect", where tiny changes snowball. This result shows chaos is not the only route: even a calm, rule-bound system can hide its ending until it has organised itself enough to reveal it. For students, it is a reminder that patterns are not always visible at the start of a problem — sometimes you have to work through the steps before the shape of the answer appears. If your child enjoys puzzles like this, you can find a maths tutor to explore patterns and prediction with them, or post a tutoring request to compare tutors.

Book your first lesson in maths and learn how patterns, rules and prediction really work: Book your first lesson

Sources:

  • Koopmans, L., Kay, E. M. and Youk, H., "Predictability can be dynamically constructed in deterministic systems", Nature Communications (11 September 2026), DOI: 10.1038/s41467-026-77737-0: https://doi.org/10.1038/s41467-026-77737-0
  • University of Illinois Physics, "The emergence of predictability in an unpredictable system", 6 November 2026: https://physics.illinois.edu/news/hyun-youk-emergence-of-predictability
  • Newswise / University of Illinois Grainger College of Engineering, "The Emergence of Predictability in an Unpredictable System", 6 November 2026: https://www.newswise.com/articles/the-emergence-of-predictability-in-an-unpredictable-system

Curious about maths? Post a request on London Tutors and compare verified maths tutors — online, in person or in a group class.

Ready to find the right tutor?

Browse verified tutors, compare profiles, and book a trial lesson — free to sign up.

Find a tutor — sign up free

Related guides