The Juggling Puzzle That Took 55 Years to Solve
Ronald Graham was a renowned mathematician, a former president of the American Mathematical Society — and a serious juggler, who also served as president of the International Jugglers' Association. He loved patterns, in numbers and in the air. In 1971 he posed a question that some mathematicians now think may have been inspired by juggling. It took 55 years, four papers and several young mathematicians to answer it.
The question is easy to state. Start with a set of different integers, not including zero. Can you always rearrange them so that, as you add them up one by one — first two, then first three, then first four, and so on — every running total, or "partial sum", turns out different?
In juggling terms, imagine each ball stays in the air for a different length of time. The question becomes: can you always find an order to throw them in so that no two balls land on the same beat?
When ordinary numbers behave, and clock numbers do not
If all the numbers are positive, the answer is obviously yes: the running totals only ever grow. If the set mixes positive and negative numbers, the answer was also known to be yes. The hard case is "clock arithmetic" — numbers wrapped around a clock face so they repeat after some prime number, p. On a 7-hour clock, for example, 3 + 4 gives 0, because 7 is the same as 0.
Graham conjectured that the answer should still be yes in clock arithmetic: for any set of non-zero numbers, and any prime p, a rearrangement with all-different partial sums always exists. He reasoned that even under rigid rules, there is usually enough flexibility to build a special pattern — much as a valid Sudoku board can usually be found despite its many rules. For decades, nobody could prove him right.
Four papers, one shared idea: start at random
The solution came in pieces, each handling sets of a different size compared with p. Alp Müyesser of the University of Oxford and Alexey Pokrovskiy of University College London tackled very large sets, where a random ordering gets most of the way there, with a few specially chosen numbers held back to fix any sequence that adds up to zero. Noah Kravitz and Benjamin Bedert, both of Oxford, solved the case of very small sets. A further paper in August 2025 extended the large-set approach. Between the small and large cases, a gap remained: medium-sized sets, containing roughly half as many numbers as p.
That gap was closed by Lisa Sauermann of the University of Bonn and Huy Tuan Pham of the University of Chicago. Their final paper appeared in February 2026, officially closing the problem. Like their predecessors, they began with a random ordering and a procedure to fix any zero-sum sequence by swapping in another number. The difficulty was proving that the fixing procedure would not itself fail — for example, by running into a bad sequence at the very end, with nothing left to swap in.
Using Fourier analysis, which rewrites functions as sums of simple waves, they showed that no one sum is especially likely to appear when random sets of numbers are added up. From that, they estimated the chance of each kind of failure and showed the total chance of failure was less than 100%. That was enough: a good rearrangement must exist. In fact, they showed a random ordering could be rearranged to remove all bad events at least 90% of the time.
One honest limit
The four papers together prove the conjecture for sets of all sizes — but only when p is very large. Nobody has calculated its exact value; think along the lines of 10 raised to the power 100. Resolving the conjecture for every p remains open, and a real juggling routine based on it would have to be far too long to perform. The result is about the structure of numbers, not a trick to learn with three balls.
Still, the story is a lovely lesson in how mathematics advances: a simple question, a long silence, then collaboration across fields, with randomness — deliberately shuffling, then carefully repairing — doing the heavy lifting. As Noga Alon of Princeton University put it, it was "the power of collaboration, the power of the young generation, the power of probabilistic methods" that solved it.
If your child enjoys puzzles where the rules look simple but the answer takes real thought, a first lesson in maths can start exactly there — with clock arithmetic, patterns, and the question of what "always" really means. Book your first lesson
Sources:
- Quanta Magazine, "Mathematicians Harness Randomness To Crack a 55-Year-Old Conjecture", 28 September 2026: https://www.quantamagazine.org/mathematicians-harness-randomness-to-crack-a-55-year-old-conjecture-20260928/
London Tutors — post your tutoring request and we will match you with a verified tutor.