In clinical practice, visual observation is often used to determine functional impairment and to evaluate treatment following a knee injury. The aim of this study was to evaluate the reliability and validity of observational assessments of knee movement pattern quality during five functional tests in subjects with an anterior cruciate ligament (ACL) injury.Twelve ACL-injured men, mean age 40 years, were video filmed before and after 12 weeks of knee-specific training when performing five different functional tests: walking, knee bending, step activity, crossover hop on one leg and one-leg hop. The videos were observed by four physiotherapists, and the knee movement pattern quality, a feature of the loading strategy of the lower extremity, was scored on an 11-point rating scale. To assess the criterion validity, the observational rating was correlated with the maximum knee flexion angle in landing during the crossover hop determined by a three-dimensional motion analysis system (VICON).Inter-observer agreement between the four physiotherapists was moderate to good, ICC(1,2) 0.57-0.76 for the four test situations. Fair to good correlations were obtained between the observers' assessment and knee flexion angle, r = 0.37-0.61. The crossover hop test or one-leg hop test was ranked as the most useful test in 172 of 192 occasions (90%) when assessing knee function.The moderate to good inter-observer reliability and the moderate criterion validity found indicate that the knee movement pattern quality in ACL-injured subjects can be determined by visual observation of more demanding functional tests such as crossover hop on one leg and one-leg hop for distance.