Delay of gratification tasks have an impressive predictive value for various outcomes and are designed to measure self-regulation. Since many behavioural and psychological conditions in children are related to limitations in self-regulation, the extent to which delay tasks can be used as a screening for the detection of psychopathology is examined. Children from the general population (non-clinical sample; N = 1498; 51% girls) participated in delay tasks at the ages of 3 and 5. Parents rated ADHD and conduct problems when children were age 5 and 6, which we classified using cut-offs. Delay at age 3 was related to ADHD at age 5 (OR = 1.84) and conduct at age 6 (OR = 2.61). The results showed high specificity (77%-78%) and high negative predictive values (95%-98%), correctly identifying children below the SDQ cut-off, but low sensitivity (27%-42%), making the task unsuitable as a screening tool for children with an increased likelihood of developing psychopathology. These results were aggravated when only the first 20 s were considered, showing better specificity but worse sensitivity values.