Some finitely additive probability: Random walks (Q1366724)

Using techniques of finitely additive probability theory by Ramakrishnan and Karandikar, and some properties of strategic measures by Purves and Sudderth, the authors prove some results on random walks in this more general setting. The main result is the following: given a random walk on the real line, determined by the finitely additive probability \(\gamma\), the random walk drifts either in the direction of \(+\infty\) or of \(-\infty\), according to whether \(\gamma\) puts mass either at \(+\infty\) or at \(-\infty\); while if \(\gamma\) puts mass at both \(+\infty\) and \(-\infty\), then the random walk oscillates in both directions. However, if \(\gamma\) is tight, then the random walk behaves like the usual random walk determined by a countably additive probability.