The determination of hashes based on biometric data is a recent topic in biometrics as it allows to handle biometric templates in a privacy manner. Applications could be the generation of biometric templates for authentication or of cryptographic keys based on biometric traits. Depending on the application, there are different requirements with regard to possible errors. On one side, authentication performance based on biometric hashes as feature representation can be measured by common biometric error rates such as FNMR, FMR and EER. Thus, generated hashes for each single person have to be only similar in a certain degree, but not necessarily identical. On the other side, biometric hashes for cryptographic issues have to be identical and unique for each individual, although measured data from same person differs from one acquisition to next, or data from different people may be similar. Therefore, we suggest three measures to estimate the reproducibility performance of biometric hash ...