Kravitz’s denominator conjecture for the Lonely Runner spectrum

For every sextuple of distinct positive integers v1,…,v6v_1,\ldots,v_6, define ML⁡(v1,…,v6)=max⁡t∈Rmin⁡1≤i≤6∥tvi∥\operatorname{ML}(v_1,\ldots,v_6)=\max_{t\in\mathbb{R}}\min_{1\leq i\leq 6}\|t v_i\|, where ∥x∥=min⁡m∈Z∣x−m∣\|x\|=\min_{m\in\mathbb{Z}}|x-m|. If ML⁡(v1,…,v6)=p/q\operatorname{ML}(v_1,\ldots,v_6)=p/q in lowest terms and ML⁡(v1,…,v6)<1/6\operatorname{ML}(v_1,\ldots,v_6)<1/6, must q=6p+kq=6p+k for some k∈{1,3}k\in\{1,3\}? In particular, must every such denominator qq be odd?

References

Primary source

arXiv

Progress summary

Refreshed
Claimed progress

A new paper explains most of the six-speed denominator pattern, but finitely many cases remain unresolved and the original conjecture already has counterexamples.

Kravitz’s conjecture proposed a specific rational-denominator pattern for values in the Lonely Runner spectrum. It was proved in low dimensions but later disproved in general, including by a six-speed example; the current question concerns the remaining six-speed pattern and its finite exceptions.

Known results

  • Kravitz proved the original spectrum conjecture for n=2,3n=2,3 and obtained partial results for n=4,6n=4,6.
  • Fan and Sun exhibited counterexamples for n=4n=4 and n=6n=6, including ML⁡(5,6,11,17,23,28)=8/51\operatorname{ML}(5,6,11,17,23,28)=8/51 (2023).
  • The amended pattern allows ML⁡(v1,…,vn)=s/(ns+k)\operatorname{ML}(v_1,\ldots,v_n)=s/(ns+k) with k≤nk\leq n; computations observed k∈{1,3}k\in\{1,3\} for n=6n=6, but the explanation was previously conjectural.

September 3, 2026 six-speed theorem

A new paper determines three infinite two-parameter families, characterizes when k=3k=3 occurs, and reports an exhaustive search of more than two billion sextuples with speeds at most 110110 finding no exception. It establishes the odd-denominator pattern apart from finitely many six-speed and five-speed tuples, but does not settle those remaining tuples.

Current status (as of September 2026): The original universal conjecture is false, while the revised six-speed denominator pattern has claimed progress but remains unverified and leaves finitely many tuples open.

Sources

Solutions 0

No solutions have been posted yet.