Publication View

A phase transition in random coin tossing (2007)

Abstract
Suppose that a coin with bias θ is tossed at renewal times of a renewal process, and a fair coin is tossed at all other times. Let µθ be the distribution of the observed sequence of coin tosses, and let un denote the chance of a renewal at time n. Harris and Keane in [10] showed that if n=1 u2 n = ∞, then µθ and µ0 are singular, while if � ∞ n=1 u2 n < ∞ and θ is small enough, then µθ is absolutely continuous with respect to µ0. They conjectured that absolute continuity should not depend on θ, but only on the square-summability of {un}. We show that in fact the power law governing the decay of {un} is crucial, and for some renewal sequences {un}, there is a phase transition at a critical parameter θc ∈ (0, 1): for |θ | < θc the measures µθ and µ0 are mutually absolutely continuous, but for |θ |> θc, they are singular. We also prove that when un = O(n −1), the measures µθ for θ ∈ [−1, 1] are all mutually absolutely continuous.

Publication details
Download http://citeseerx.ist.psu.edu/viewdoc/summary?doi=?doi=10.1.1.22.759
Source http://www.math.uconn.edu/~levin/papers/coin.pdf
Contributors CiteSeerX
Repository CiteSeerX - Scientific Literature Digital Library and Search Engine (United States)
Type text
Language English
Relation 10.1.1.56.2276, 10.1.1.28.5338