IBM Research releases new optimization puzzle
IBM Research has released its August 2026 Ponder This challenge, "The Wheel of Buttons," inviting researchers to solve a complex optimization puzzle with implications for algorithmic design.

IBM Research has published its August 2026 Ponder This challenge, titled The Wheel of Buttons, presenting a new mathematical puzzle designed to test algorithmic optimization. The challenge describes a game show featuring a symmetric wheel with N buttons, each of which can be toggled between on and off states. Because there are no indicators showing whether a button is currently active, and the wheel spins invisibly between rounds, contestants must formulate a blind strategy to guarantee a win by turning all buttons on in the minimum number of rounds, even under adversarial conditions.
The game represents strategies as sequences of natural numbers where rounds are separated by 0. For a simple wheel where N equals 2, the optimal solution sum of the pressed button indices is 7. If a player were to press button 2 instead of button 1 during the middle round, this sum would increase to 8. The core of the August challenge is to calculate the optimal solution sum for N equals 8, while a special bonus is offered for solving the highly complex N equals 64 variant. Several researchers solved the puzzle immediately after its release, starting with Daniel Chong Jyh Tar on July 31, 2026, at 4:54 PM IDT, and followed by others like Jean-Francois Hermant and Rethna Pulikkoonattu by August 2, 2026.
For computer science and AI practitioners, these monthly challenges are more than mere brainteasers. The Wheel of Buttons is a classic example of state estimation and control under extreme uncertainty, resembling problems in blind search, cryptography, and self-stabilizing distributed systems. Developing an optimal strategy for such puzzles helps researchers design robust protocols that can bring decentralized networks into a desired state without relying on continuous feedback or assuming benevolent environments.
This is our own summary of reporting by IBM Research AI



