Will a random walk hit every point infinitely often?

Click For Summary

Discussion Overview

The discussion centers around the behavior of a random walk, particularly whether it will hit every point infinitely often or cross every bound infinitely often. The focus includes theoretical aspects of random walks, step sizes, and their distributions.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • Some participants propose that a random walk with steps having mean 0 and values greater than ε with positive probability will be unbounded and cross bounds infinitely many times in 1D.
  • Others argue that if the set of allowed step sizes includes all real numbers, it is impossible to hit every point due to the uncountable nature of points versus the countable nature of steps.
  • A participant later clarifies that the intention was to discuss crossing every bound infinitely often rather than hitting every point.
  • It is mentioned that for a step size of 1, there is a well-known result indicating that even a small positive fraction of steps being 1 (with the rest being zero) leads to crossing bounds infinitely often, which can be scaled to ε.
  • Another participant inquires about the name of the result related to crossing bounds infinitely often.
  • A reference to Wikipedia is made, indicating that the result is known by several names, including the level-crossing phenomenon, recurrence, or the gambler's ruin.

Areas of Agreement / Disagreement

Participants express differing views on the implications of step sizes and distributions, with no consensus reached on whether every point can be hit or only bounds crossed infinitely often.

Contextual Notes

There are limitations regarding the definitions of step sizes and their distributions, as well as the implications of countability versus uncountability in the context of random walks.

Who May Find This Useful

This discussion may be of interest to those studying stochastic processes, random walks, or mathematical theories related to probability and recurrence.

hyurnat4
Messages
22
Reaction score
0
Well, not quite a random walk. The steps aren't necessarily ±1, but they have mean 0 and will take values >ε with positive probability. It seems intuitive that such a process will be unbounded and will cross this bound infinitely many times (in 1D). Does anyone know of a result that says this?
 
Last edited:
Physics news on Phys.org
What is the set of allowed step sizes, and what is its distribution? If you consider all real numbers, it is impossible to hit every point as there are uncountable points and countable steps.
 
  • Like
Likes   Reactions: hyurnat4
mfb said:
What is the set of allowed step sizes, and what is its distribution? If you consider all real numbers, it is impossible to hit every point as there are uncountable points and countable steps.

Sorry, it was a bad title. I meant 'cross every bound' infinitely often.
 
For a step size of 1 it is a well-known result, even if just a small positive fraction of all steps is 1 and the rest is zero. This is trivial to scale to ε. All you have to do is to show that additional small steps instead of zero and additional larger steps instead of ε don't change that result.
 
  • Like
Likes   Reactions: hyurnat4
mfb said:
For a step size of 1 it is a well-known result, even if just a small positive fraction of all steps is 1 and the rest is zero. This is trivial to scale to ε. All you have to do is to show that additional small steps instead of zero and additional larger steps instead of ε don't change that result.

Ah thank you. Does the result have a name?
 
According to wikipedia:
This result has many names: the level-crossing phenomenon, recurrence or the gambler's ruin.
 

Similar threads

  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 4 ·
Replies
4
Views
5K
  • · Replies 9 ·
Replies
9
Views
4K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 66 ·
3
Replies
66
Views
7K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 8 ·
Replies
8
Views
7K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 3 ·
Replies
3
Views
9K