Answers are grounded in this paper and cite their sources.
Let E be the middle-thirds Cantor set and, for an integer b≥ 2, let νb be the measure on E obtained by pushing forward the harmonic measure of simple random walk on the tree of b-adic intervals that meet E. The measure ν₃ is the Cantor measure. Lyons and Peres asked whether ν₂ and ν₃ are mutually singular. We show that they are. The proof compares the conditional splits of the two measures on dyadic cells through Hellinger affinities. At one explicit dyadic cell, whose intersection with E is a single ternary cylinder, the Cantor measure splits evenly while the harmonic measure, computed as an electrical current, does not; an exact computation with rational arithmetic bounds the harmonic split away from $1/2$, and the discrepancy persists for all nearby positions and scales of the cylinder. A Fourier argument resting on Stewart's theorem on digit changes in the bases $2$ and $3$ then shows that, if the two measures were not singular, such cells would carry a fixed share of their overlap at infinitely many levels, which is incompatible with the summability of the Hellinger losses.
No takes yet. Share an insight, caveat, or question.
Yitzchak Shmalo (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: