← Back to Paradoxes

The Moving Sofa Problem

What's the largest shape that can navigate an L-shaped corner?

Solved November 2024
The Problem: You're moving a sofa through an L-shaped hallway. What's the largest possible sofa that can make it around the corner? After 58 years, mathematicians finally have the answer: 2.2195...

A Deceptively Simple Question

Everyone who has ever moved furniture knows the frustration: that perfect couch that just won't fit around the corner. In 1966, mathematician Leo Moser transformed this everyday annoyance into one of geometry's most stubborn problems.

The rules are simple: an L-shaped corridor with unit width (1 meter). A rigid 2D shape must slide through, rotating as needed, without overlapping the walls. What's the maximum area this shape can have?

Surely, with modern mathematics and computers, this should be trivial? Yet for nearly six decades, the best minds in geometry could only narrow down the answer to a range. Until November 2024.

Watch It Move

Interactive Sofa Navigator
2.2195
Area
0%
Progress
Rotation
Ready
Status

A Brief History of Sofa Optimization

1966
Leo Moser poses the problem. Initial bound: a semicircle has area π/2 ≈ 1.57
1968
John Hammersley designs a "telephone" shape with a semicircular bite, achieving area π/2 + 2/π ≈ 2.2074
1992
Joseph Gerver constructs a sofa from 18 analytic curves, area ≈ 2.2195. Conjectures it's optimal.
2018
Kallus and Romik prove upper bound of 2.37—no sofa can be larger than this.
November 2024
Jineon Baek proves Gerver's sofa is optimal! The 58-year quest is complete.

Compare the Shapes

Rectangle

1.00

The naive solution. A 1×1 square just barely fits.

Semicircle

1.57

π/2—a natural first improvement using circular motion.

Hammersley

2.21

A semicircle with a bite taken out. Major breakthrough.

Gerver (Optimal)

2.2195

18 precise curves. Proven optimal in 2024.

Why Was This So Hard?

The difficulty lies in the continuous nature of the problem. Unlike discrete puzzles with finite possibilities, the sofa can be any shape. Every curve, every indentation, every bulge must be considered.

The Core Insight: The sofa must rotate as it navigates the corner. This means different parts of the sofa sweep different paths. The optimal shape balances two competing goals: maximize area while ensuring every point can complete its required trajectory.

Gerver's breakthrough was recognizing that the optimal sofa's boundary consists of 18 distinct curves, each satisfying specific differential equations. The shape has:

Area = 2.2195316688...
(approximately π/2 + 2/π + tiny optimizations)

The Ambidextrous Variant

John Conway posed a variant: what if the sofa must navigate both left AND right turns? This "ambidextrous sofa" problem has a different optimal shape—symmetric and smaller.

Dan Romik found such a sofa with area ≈ 1.64495. The constraint of bilateral navigation significantly reduces the maximum achievable area.

Real-World Implications

While purely mathematical, the moving sofa problem connects to practical concerns in:

Robotics

Path planning for autonomous vehicles and warehouse robots navigating tight spaces.

Manufacturing

Designing machinery that must be transported through factory corridors.

Architecture

Understanding furniture accessibility when designing building layouts.

The Deeper Lesson

The moving sofa problem reminds us that simple questions can hide profound complexity. A child can understand the problem in seconds. Proving the answer took humanity 58 years.

"The simplest-sounding problems are often the hardest to solve—because simplicity in statement doesn't imply simplicity in structure."