Offending patterns in stranger sexual offences against women were analysed to identify crime scripts, and the predictive factors and contexts involved in unsolved cases. The aim of the study was to explore what patterns could be found in sexual offences against women for unsolved offences. Crime script analysis, hierarchical logistic regression and conjunctive analysis were used to analyse data from 542 cases from the Violent/ Sexual Crime Database (VSCD) based on police data. The sample was limited to sexual offences against women (16 years or above) by male stranger offenders, in a hands-on sexual offence in Queensland, Australia. An overall crime script was extracted, as well as script tracks for solved and unsolved cases. Differences were found in the location/ setting, offender approach method and offender-victim interaction stages. Several variables were found to be statistically significant in predicting whether a crime was solved or not: offender consumption of alcohol or drugs prior to the offence, vehicle used, bystander presence at any location, minimal offender force used, and if the victim was attacked while walking at the time of initial contact with the offender. After controlling for the presence of forensic evidence and case linkages within the sample, the two strongest predictors were offender use of minimal force and offender prior use of alcohol or drugs. Overall two themes emerged from the findings that impacted whether a case became solved or unsolved: criminal effectiveness and the level of personal engagement between the offender and victim. The concept of the travelling victim was also identified in relation to unsolved crimes. Prevention implications are discussed.