Likelihood ratio (LR) tests are provided for the three alternatives to DNA identity: exclusion, coincidence, and kinship. The coincidence test uses the radius of coalescence to conserve the observed frequency of single band phenotypes. Genotype probabilities under kinship are derived for mating groups, specified relatives, and structured populations; and unbiased estimates of the genetic parameters are provided. The LR is made robust to gene frequency errors by specifying the mean matching probability, and the tolerable loss of information this entails is determined by LR theory. This straightforward application of the seminal work of Jerzy Neyman and Sewall Wright strongly supports the use of LRs and kinship for presentation of DNA evidence by expert witnesses and committees.