Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Cover times for Brownian motion and random walks in two dimensions
radius_distribution: Derives the exponential radius law with exact integer and boundary conventions.
theorem_1_1: The time simple random walk on the discrete torus (Z/nZ)^2 takes to visit every site, divided by (n log n)^2, tends to 4/pi in probability, proving Aldous's conjecture.
theorem_1_2: For Brownian motion on the two-dimensional torus, the time to come within epsilon of every point, divided by (log epsilon)^2, tends to 2/pi almost surely as epsilon tends to 0.
theorem_1_3: For Brownian motion on a smooth compact connected two-dimensional Riemannian manifold without boundary and of area A, the epsilon-covering time divided by (log epsilon)^2 tends to 2A/pi almost surely.
theorem_1_4: States the sharp distributional law for covering a disc centered at the origin.
Amir Dembo, Yuval Peres, Jay Rosen, and Ofer Zeitouni, Cover times for Brownian motion and random walks in two dimensions, Annals of Mathematics 160 (2004), 433–464, DOI 10.4007/annals.2004.160.433.
Source versions
The copy read for this card is the published PDF, the 32-page publisher copy. Printed p. 433 is PDF p. 1. The publisher record identifies the volume, issue and page range.
An earlier copy read for this card is the 30-page arXiv v2. It identifies arXiv:math/0107191v2, 27 November 2003. Its PDF pagination is not the published pagination, and no line-by-line proof equivalence between the versions is asserted. All result citations below use the published version. The published PDF prints no notice on any of its 32 pages; the journal's article page, which links it as a free download, shows only the footer "Copyright © 2026 Annals of Mathematics" and names no license (https://annals.math.princeton.edu/2004/160-2/p02, read 2026-10-02), every other right reserved. For the arXiv copy, the arXiv record carries no license field, so arXiv's assumed license applies (arXiv:math/0107191), every other right reserved.
Problem connection
Theorem 1.4, printed p. 436, settles the Kesten–Révész distributional conjecture for the time needed by a planar simple random walk to cover a disc centered at the origin. Its exact inversion gives the radius formulation of Problem 1164. The original 1999 question uses a random radius and an exponential cumulative distribution.
The radius deduction is fully written relative to the stated cover-time input. Theorem 1.4 itself remains a precise statement with a proof pointer: its complete same-paper chain has not yet been reconstructed here.
Main results
- Theorem 1.1, printed p. 434: the cover time of the lattice torus by simple random walk satisfies in probability (Aldous's conjecture). Section 4, printed pp. 446–447, proves the lower half from Theorem 1.2 by strong approximation; the upper half is cited from Aldous and Fill.
- Theorem 1.2, printed p. 435: for Brownian motion on the flat torus , the -covering time satisfies almost surely as . The upper bound is Section 2 (pp. 437–443), the lower bound Section 3 (pp. 443–446), with estimates from Sections 6 and 7 (pp. 452–458).
- Theorem 1.3, printed p. 435: on a smooth compact connected two-dimensional Riemannian manifold without boundary and of area , the limit is almost surely; Section 8, pp. 459–461.
- Theorem 1.4, printed p. 436: the disc-cover law above; Section 5, pp. 447–452.
The common method counts excursions between concentric circles on many scales at once and runs a second-moment argument for uncovered points while controlling the dependence on excursion endpoints. The torus walk of Theorem 1.1 is confined to ; Theorem 1.4 concerns the walk on all of . Section 9 (pp. 461–462) adds Corollary 9.1 on the largest unvisited disc of the torus walk, recorded on the Theorem 1.1 page, and open problems.
Read status. Claims checked: Theorems 1.1–1.4 and the definitions they use (pp. 434–437, 447 and 459) were read clause by clause on the printed pages. The proofs were read for structure only; the radius deduction is this corpus's own argument from the stated Theorem 1.4. Nothing here is independently reviewed.
Bears on. #1164: Theorem 1.4 gives the limit law of the cover time of an origin-centered disc; the corpus's radius deduction inverts it to the law , , for the covered radius , which implies the two-sided comparison in probability that the problem's corrected Statement asks about. Theorems 1.1–1.3 do not bear on the problem.
No file of this source is held: no license on record permits its redistribution, and the card cites the edition it names above.