Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Setting (p. 137). The walk is the symmetric nearest-neighbor walk on started at the origin.
Theorem 7 (p. 151). Let be a monotonic function increasing to , and let be the event that the planar walk does not return to the origin between times and . Then is or according as
converges or diverges. The series runs over the doubly exponential sequence , the sequence of the proof. The print writes the growth hypothesis as "increases to as " [sic], with the variable for .
The paper motivates the theorem (p. 151) as the question of how long the gaps between returns can be: for which monotone the interval contains a return for all but finitely many .
Theorem 7A (p. 153). The paper states without proof the line analogue, saying it can be proved by similar methods: for the walk on , with as above and the event of no return to the origin between and , is or according as converges or diverges.
Source. P. Erdős and S. J. Taylor, Some problems concerning the structure of random walk paths, Acta Math. Acad. Sci. Hungar. 11 (1960), 137--162: the walk on p. 137, (2.16) on p. 141, Theorem 7 and (4.8)--(4.9) on p. 151, the proof on pp. 151--153, Theorem 7A on p. 153. The edition read is identified on the source card.
Read depth. Claims checked: Theorems 7 and 7A were read clause by clause on the printed pages. The proof on pp. 151--153 was read for the pointer below and not checked step by step. Nothing here is independently reviewed.
Proof pointer
Pages 151--153. The paper's estimate (2.16) (p. 141) for the probability that a planar walk started at distance avoids the origin for steps gives, from the position at time , two-sided bounds for large ((4.8)--(4.9), p. 151). Convergence of the series gives the zero case by Borel--Cantelli along , with replaced by to cover the intermediate . For divergence the paper bounds the overlap in two cases ((4.14)--(4.15), p. 153), so that disjointified events carry at least half the probability, and concludes with the zero-one law.
Dependencies
The planar avoidance estimate (2.16) of the same paper (p. 141), which rests on (2.5) and the local estimates (2.9)--(2.10).
Bears on
No problem page of this corpus.