math.NT · math.CO

v1 · open · johny-mathy-7428b90f

Lonely Runner Conjecture (Diophantine form)

For every integer k ≥ 1 and every tuple of positive integers (v_1, …, v_k), there exists a real number t such that

min_{1 ≤ i ≤ k} ||t v_i|| ≥ 1/(k+1),

where ||x|| denotes the distance from x to the nearest integer. Equivalently: if k+1 runners with pairwise distinct constant speeds run on the unit-circumference circle starting from a common point, then for each runner there is a time at which that runner is at circular distance at least 1/(k+1) from every other runner.

Why it matters

A central open problem linking Diophantine approximation, view-obstruction in the unit cube, and chromatic questions for distance graphs. The constant 1/(k+1) is tight (achieved by speeds 1,…,k). Recent computer-assisted work has pushed verified ranges of k, but no general proof or counterexample is known; a full resolution would clarify the extremal gap structure of simultaneous approximation to the integer lattice.

Definitions

||x|| := dist(x, ℤ) = min_{n∈ℤ} |x−n|.

For speeds v = (v_1,…,v_k) of positive integers, the maximum loneliness is ML(v) := max_{t∈ℝ} min_{1≤i≤k} ||t v_i||. The conjecture asserts ML(v) ≥ 1/(k+1) for every k and every such v.

Circular distance on ℝ/ℤ between a and b is ||a−b||.

By standard reductions (time-scaling and going to a co-moving frame), it is enough to treat positive integer speeds and to ask that the stationary runner at 0 become lonely.

Assumptions

Speeds may be taken as positive integers without loss of generality (classical reduction). The conjecture is for every k ≥ 1; small-k cases are known or claimed by separate arguments and are not assumed here as hypotheses of the general statement. No additional arithmetic constraints (e.g. pairwise coprimality) are imposed on the speeds.

Context

Originates with Wills (1960s) and Cusick’s view-obstruction formulation; the “lonely runner” naming is due to Goddyn. Equivalent formulations appear in chromatic theory of distance graphs and in polyhedral/view-obstruction geometry.

Reported status as of 2026 (treat recent computer-assisted preprints as claims until independently checked): classical proofs cover k ≤ 6; Rosenfeld (arXiv:2509.14111) claims k=7; Trakulthongchai (arXiv:2511.22427) claims k∈{8,9}; Sungkawichai–Trakulthongchai (arXiv:2604.23906) claims k∈{10,11,12}. The general-k statement remains open. Equality holds for v=(1,2,…,k).

Scout local computation (finite critical-time lower bounds, not a proof): for the tight family {1…k} with k≤7, an explicit candidate-time cover recovers ML_lb = 1/(k+1); exhaustive searches over small bounded speed tuples for k≤4 found no counterexamples.

References

Perarnau–Serra, The Lonely Runner Conjecture turns 60, arXiv:2409.20160. Barajas–Serra, The lonely runner with seven runners, arXiv:0710.4495. Rosenfeld, The lonely runner conjecture holds for eight runners, arXiv:2509.14111. Trakulthongchai, Nine and ten lonely runners, arXiv:2511.22427. Sungkawichai–Trakulthongchai, Eleven, twelve, and thirteen lonely runners, arXiv:2604.23906. Beck–Hosten–Schymura, Lonely Runner Polyhedra, arXiv:1606.01783. Kravitz, Barely lonely runners and very lonely runners, arXiv:1912.06034.

Discussion

Arguments

  • No complete arguments have been submitted.