Conflict tasks play a central role in the study of self-control. These tasks feature a condition assumed to demand top-down control and a matched condition where control demands are assumed to be absent, and individual differences in control ability are indexed by subtracting measures of performance (e.g., reaction time) across these conditions. Subtraction-based metrics of top-down control have recently been criticized for having low test-retest reliability, weak intercorrelations across conceptually similar tasks, and weak relationships with self-report measures of self-control. Concurrently, there is growing evidence that task-general cognitive efficiency, indexed by the drift rate parameter of the diffusion model (Ratcliff, 1978), constitutes a cohesive, reliable individual difference dimension. However, no previous studies have examined the measurement properties of subtraction metrics of top-down control as compared to drift rate in the same sample, or compared their respective associations with self-report measures. In this re-analysis of open data drawn from a large recent study (Eisenberg et al., 2019; N=522), we find that subtraction metrics fail to form cohesive latent factors, that the resulting factors have poor test-retest reliability, and that they exhibit tenuous connections to questionnaire measures of self-control. In contrast, cognitive efficiency measures from the same tasks form a robust, reliable latent factor that shows moderate associations with self-control. Importantly, this latent cognitive efficiency variable is constructed from conditions that both were, and were not, previously assumed to index control. These findings invite a reconceptualization of subtraction-based tasks, pointing to task-general efficiency as a central individual difference dimension relevant to self-regulation.