BACKGROUND: Many studies address diagnosing concussions, but few look at predicting prognosis. A previous discriminant function analysis showed that symptom clusters derived from the Post-Concussion Symptom Scale and Immediate Postconcussion Assessment and Cognitive Testing composite scores used together improved predictions of protracted recovery after a sports-related concussion. OBJECTIVE: To determine cutoff scores in neurocognitive and Post-Concussion Symptom Scale symptom cluster scores when classifying protracted recovery in concussed athletes. METHODS: 108 male high school football athletes completed a computer-based neurocognitive test battery (Immediate Postconcussion Assessment and Cognitive Testing) within a median of 2 days after injury. Patients completed graded exertional protocols requiring athletes to be symptom free at rest and during increasing levels of activity and had recovery of neurocognitive scores before return to play. After return to play, athletes were classified as protracted recovery (>14 days, n = 58) or short-recovery (=14 days, n = 50). Receiver-operating characteristic curves analyzed each of the neurocognitive (verbal, visual, processing speed, and reaction time) and symptom cluster (migraine, cognitive, sleep, and neuropsychiatric) scores. RESULTS: Cutoffs for migraine cluster, cognitive cluster, visual memory, and processing speed were statistically significant. Cutoffs at 75%, 80%, and 85% sensitivity to predict protracted recovery for the migraine symptom cluster were 15 or greater, 18, 20; cognitive symptom cluster 18 or greater, 19, 22; visual memory 48 or less, 46, 44.5; and processing speed 24.5 or less, 23.46, 22.5, respectively. Eighty-percent sensitivity indicates that the corresponding cutoff correctly identify 80% of concussed athletes requiring protracted recovery. CONCLUSION: Specific cutoffs may help to set numerical thresholds for clinicians to predict which concussed athletes will have a protracted recovery.