The best test to assess dual task (DT)-related falls' risk is not known. The aim of this study was to investigate differences between community-dwelling fallers and non-fallers on a variety of simple task combinations. Twenty-seven adults, aged 65 years or older, took part. Forty-eight DT tests and one triple task (TT) test were conducted. Fallers had longer walking time when avoiding a moving obstacle and performing a motor task and longer walking time when triple tasking, as suggested by a measure of proportionate difference between single and DT/TT conditions (p = 0.014 and 0.044, respectively). The absolute difference in accuracy on a visuospatial task suggested that fallers were more accurate than non-fallers when combined with walking with turns and when triple tasking (p = 0.048 and 0.030, respectively). Fallers were less accurate in naming animals than non-fallers when combined with a bending task (p = 0.009). These results indicate that fallers might prioritise tasks based on perceived risk, which highlights the importance of task selection when designing tests. Despite the small sample size, the data suggest that a TT test could be used to assess risk of falling. However, this needs to be confirmed with larger prediction studies.