A common strategy for testing for linkage without posing a disease model is to test for increased marker similarity among the affected pedigree members. We developed a simulation based statistic, SimIBD, which measures marker similarity in terms of identity-by-descent (IBD) when it can be determined whether or not the alleles are IBD, and in terms of the probability of the alleles being IBD when it cannot be determined. The SimIBD statistic is not only more powerful than its precursor, the affected-pedigree-member (APM) method, but it is also less sensitive to the misspecification of marker allele frequencies.