1: \begin{abstract}

2: Metropolis algorithms for approximate sampling of probability measures on infinite dimensional Hilbert spaces are  considered and a generalization of the preconditioned Crank-Nicolson (pCN) proposal is introduced. 

3: The new proposal is able to incorporate information on the measure of interest.   

4: A numerical simulation of a Bayesian inverse problem indicates that a Metropolis algorithm with such a proposal 

5: performs independently of the state space dimension and the variance of the observational noise.

6: Moreover, a qualitative convergence result is provided by a comparison argument for spectral gaps.

7: In particular, it is shown that the generalization inherits geometric convergence

8: from the Metropolis algorithm with pCN proposal.

9: \end{abstract}