An Error in the Mixed Adversary Protocol by Fitzi, Hirt and Maurer

Abstract

We point out an error in the protocol for mixed adversaries and zero error from the Crypto 98 paper by Fitzi, Hirt and Maurer. We show that the protocol only works under a stronger requirement on the adversary than the one claimed. Hence the bound on the adversary's corruption capability given there is not tight. Subsequent work has shown, however, a new bound which is indeed tight.