Background: The Oxfordshire Community Stroke Project (OCSP) and Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classifications are widely used for the assessment of major ischaemic stroke. We explored their intra- and inter-observer reliability in the classification of outpatient minor stroke. Methods: Four physicians of differing seniority and training backgrounds classified minor stroke using clinical data from 90 patients. Results: For both the OCSP and TOAST classifications, the intra-observer reliability varied from moderate to excellent (κ = 0.48–0.83). The inter-observer reliability was good (κ = 0.64) for the OCSP and moderate (κ = 0.42) for the TOAST. Thus, neither classification was consistently reliable. Conclusions: Our results may reflect the limited validity of these classifications in a typical minor stroke outpatient population and variable observer expertise.