Purpose: To compare critical speed (CS) derived from all-out testing (AOT) for linear and shuttle running with metrics from a graded exercise test, the Yo-Yo Intermittent Recovery Test  Level 1 (YYIR1), and estimation of an 800-m-shuttle time trial. Methods: Twelve male rugby players completed a graded exercise test, the YYIR1, a linear AOT, shuttle AOTs of 25 and 50 m, and an 800-m-shuttle time trial consisting of 32 × 25-m shuttles. Results: Strong linear correlations were observed between maximum oxygen uptake ($\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$) and CS (m·s⁻¹) derived from the linear AOT (3.68 [0.62], r = .90, P < .01) and 50-m-shuttle AOT (3.19 [0.26], r = .83, P < .01). Conversely, $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$ showed lower correlations with speeds evoking CS from 25-m AOT (2.86 [0.18], r = .42, P = .18) and YYIR1 (4.36 [0.11], r = .55, P = .07). The 800-m time trial (213.58 [15.84] s) was best predicted using parameters from the 25-m AOT (r = .93, SEE = 6.60 s, P < .001). Conclusions: The AOT is a valuable method of assessing performance-specific fitness, with CS from linear and 50-m-shuttle AOTs being strong predictors of $\dot{\mathrm{V}}{\mathrm{O}}_2\text{max}$, rivaling metrics from the graded exercise test. The YYIR1 offered limited utility compared with the AOT method.