For decades, one geometry puzzle that sounded simple stayed unsolved, outliving generations of researchers and countless computer experiments. A 31-year-old mathematician from South Korea has now finished it, not with supercomputers, but with a pencil, paper, and a lot of patience.
In 1966, Leo Moser, an Austrian-Canadian mathematician, asked a question that many homeowners might know. Imagine a hallway that looks like an L, with each arm being one metre wide. What is the biggest flat, stiff shape you can carry around the corner without having to lift or bend it?
People started calling this puzzle the “moving sofa problem.” It sounds like something that would happen in a student flatshare, but it quickly became a famous geometry problem. The goal is to make the area of a shape that can fit through that right angle as big as possible.
China unveils world’s first lunar clock to solve strange time dilation predicted by Einstein
In the beginning, mathematicians suggested possible answers instead of exact ones. John Hammersley came up with a shape in 1968 that could cover an area of about 2.2074 square meters. That was already bigger than a circle or rectangle could be.
Joseph Gerver, an American mathematician, pushed the limit even further in 1992. He made a complicated, curved shape with a lot of different corners and arcs. It had an area of about 2.2195 square meters and won the title of champion.
For more than thirty years, Gerver’s strange, many-curved sofa shape was the best guess, but no one could prove that it was the best.
Computer simulations showed that no bigger couch should fit. Simulations can only test certain shapes and approximations, though. They never gave a clear, mathematical “no larger shape exists.” The door was still open to the idea that a smarter design might work better.
A soldier finds a riddle that is 60 years old.
South Korea’s mandatory military service was the turning point that no one saw coming. Baek Jin-eon, a young mathematician, found the moving sofa problem while working at the National Institute for Mathematical Sciences.
The fact that it was famous and vague both caught his attention. There was no widely accepted theoretical framework or neat formula; just an awkward question that had been floating around in applied mathematics since the 1960s.
Baek’s drive came from the lack of structure. He began to think about how to make this casual puzzle into a well-defined optimisation problem with clear limits and tools.
Seven years, 119 pages, and no computers
Baek continued to work on the problem during his doctoral studies at the University of Michigan & afterward at the June E. Huh Center for Mathematical Challenges at the Korea Institute for Advanced Study.
He spent about seven years building a full proof from scratch. No numbers, no symbols, and no help from machines. Just a chain of precise arguments that had to survive every self-posed objection.
His final manuscript runs to 119 pages and argues that Gerver’s shape is not just good, but absolutely optimal.
The key breakthrough lies in how Baek formalised the puzzle. Instead of treating it as a vague question about “shapes that fit”, he encoded every possible motion of a shape through an L-shaped corridor into a rigorous mathematical setting. That allowed him to phrase the challenge as an optimisation problem with strict upper bounds.
Within this framework, Baek proves that any shape which manages to move around the corner must have an area at most equal to Gerver’s. Since Gerver’s construction reaches that limit, no sofa larger than 2.2195 square metres can succeed.
What makes this proof stand out
- It provides the first exact upper bound matching a known construction.
- It avoids reliance on numerical approximations or computer searches.
- It introduces a reusable framework for similar geometric optimisation problems.
The work has been posted on the scientific preprint server arXiv and is currently under review at the prestigious journal Annals of Mathematics. This journal publishes results that can reshape entire fields of study.
A reminder of what pure thought can do
Baek often describes his process as a cycle of hope and collapse. In interviews with Korean media, he has spoken of repeatedly believing he saw the full picture, only to find a gap and dismantle months of work.
“You keep a bit of hope, then crush it, and move forward picking ideas from the ashes,” he said, likening the process to a chain of dreams and awakenings.
At just 31, he continues to work in combinatorial geometry and optimisation, in a research environment that is rapidly boosting South Korea’s visibility in pure mathematics. His moving-sofa result showcases a style of research that feels almost old-fashioned in the age of AI and massive computation: patient, solitary, heavily conceptual.
The success also offers a subtle counter-argument to the belief that big mathematical advances now belong mainly to machines. For this particular question, at least, human persistence and abstract thinking produced the final word.
Why a “sofa problem” matters
On the surface, negotiating a hallway with a theoretical sofa looks like a curiosity. Yet the ideas behind it echo across more practical areas.
Any situation where a rigid object must move through a constrained space raises similar questions: how big can the object be, what paths are allowed, and what configuration maximises some quantity under those rules?
| Field | Related question |
|---|---|
| Robotics | Finding collision-free paths for robot arms in cramped factories. |
| Logistics | Planning how to manoeuvre large loads through warehouses or city streets. |
| Computer graphics | Computing feasible motions of objects in animation or virtual reality. |
| Manufacturing | Checking whether parts can be inserted into machines without disassembly. |
In all these settings, researchers talk about “configuration spaces”. Instead of only describing where an object is, they track every possible way it could be positioned and rotated. The moving sofa problem is a particularly tricky instance where the geometry of that configuration space becomes highly tangled.
From puzzle to broader ideas
The techniques Baek developed could shed light on other geometric puzzles that have resisted exhaustive computation. For instance, similar reasoning might help with shapes sliding through different kinds of corridors, or with three-dimensional versions where a rigid body needs to twist around obstacles.
There is also a cultural dimension. Puzzles like the moving sofa problem often appear in popular maths books, classroom discussions and outreach talks. Having a final, rigorous answer allows teachers and communicators to show students how a casual-sounding question can grow into a serious line of research involving optimisation, topology and analysis.
For readers not used to mathematical jargon, a couple of terms help frame what Baek accomplished. “Optimisation” here means finding the best possible object that satisfies all the constraints, not just a decent candidate. “Rigorous proof” means a chain of arguments so tight that every step can be justified from accepted axioms and prior results, leaving no room for guesswork or unchecked assumptions.
One way to picture the impact is to imagine architects trying to design the largest piece of modular furniture that will fit through a fixed stairwell. For years they relied on intuition, mock-ups and 3D models, unsure whether a slightly larger design might still squeak through. With an argument like Baek’s, they would finally know the absolute upper bound. No amount of extra cleverness could beat it.
The moving sofa problem now sits in that category: a 60-year-old uncertainty turned into a clean, if intricate, theorem. The riddle about a corridor, an angle and a hypothetical sofa has become a landmark in how far sheer reasoning can go when given time, focus and a question that refuses to let go.
