The Wingate all-out test (WAT) is commonly used to estimate anaerobic capabilities of athletes by using an upper or lower body cycle ergometer, however, a new test modality called elliptical all-out test (EAT) which measures activated whole-body locomotor tasks has recently been proposed. The purpose of this study was to evaluate the familiarization effects of a 30-s EAT versus WAT. Twenty male trained athletes performed pre-familiarization (Trial-I), post-familiarization (Trial-II) and retest of Trial-II (Trial-III) sessions on both cycle ergometer and elliptical trainer. Peak power (PP), average power (AP), power drop (PD) and fatigue index ratio (FI%) were analyzed using student's t-test for paired samples and correlated by intra-class correlation coefficients (ICC). Moreover, an error detection procedure was administered using data attained from illogical interrelations among 5-s segments of 30-s tests. The main results showed that there were significant familiarization effects in all mechanical power outputs obtained from Trial-I and Trial-II in both EAT (ICC = 0.49-0.55) and WAT (ICC = 0.50-0.57) performances (p <= 0.01). Significant segmental disorders were detected in power production during Trial-I of EAT, however, none existed in any of test trails in the WAT (p <= 0.001). After familiarization sessions, reliability coefficients between Trial-II and Trial-III showed moderate to strong-level agreements for both EAT (ICC = 0.74-0.91) and the WAT (ICC=0.76-0.93). Our results suggested that prior to the performance tests, combination of a well designed familiarization session with one full all-out test administration is necessary to estimate the least moderately reliable and accurate test indices for both WAT and EAT.