[lalaithion]

Ah, but this assumes a fixed unfairness; with this result, you could pretend to do the von neumann method but change starting sides on each flip, giving a biased result.