A challenge state is associated with superior performance compared to a threat state in a variety of performance domains (e.g., sport, aviation, education). However, in the challenge and threat (C/T) literature, between-subjects variability in past performance is often inconsistently controlled for. Here, we aimed to investigate the effects of C/T states on performance using two methods to control for past performance. Experiment 1 used previous performance statistics in a between-subjects design and Experiment 2 used a within-subject design. In Experiment 1, regression analysis showed that cardiovascular correlates of C/T states predicted cricket batting performance in 45 amateur cricketers. In Experiment 2, between- and within-subject analysis found that past performance was the only predictor of subsequent golf putting performance in 40 noncompetitive golfers. Taken together, the findings challenge the role that C/T states play in predicting performance under pressure after controlling for past performance.