Background: The positive role of self-efficacy in directing a wide range of health-related interventions has been well documented, including those targeting an increase in physical activity. However, rarely do researchers control the influence of past performance and past self-efficacy perception ratings when exploring the interaction of self-efficacy and performance, allowing for a refined understanding of this relationship and the unique contribution of each factor. Methods: A residualized past performance, residualized self-efficacy hierarchical regression model was used to examine the effect of prior past performance and pre-exercise self-efficacy on performance with a health-related task (12 aerobic exercise cycling sessions). Results: The previous day’s residualized performance was a significant predictor of performance, as was same-day residualized self-efficacy (P < .001). However, residualized self-efficacy became a stronger predictor over time. Conclusions: While maintaining a consistent level of moderate–vigorous physical activity over 12 exercise sessions, participants increased their ratings of task self-efficacy, explaining an increasing portion of the variance in the self-efficacy–performance relationship days 9 to 12.