The maximal oxygen uptake ( O) is the primary determinant of endurance performance in heterogeneous populations and has predictive value for clinical outcomes and all-cause mortality. Accurate and precise measurement of O requires the adherence to quality control procedures, including combustion testing and the use of standardized incremental exercise protocols with a verification phase preceded by an adequate familiarization. The data averaging strategy employed to calculate the O from the breath-by-breath data can change the O value by 4-10%. The lower the number of breaths or smaller the number of seconds included in the averaging block, the higher the calculated O value with this effect being more prominent in untrained subjects. Smaller averaging strategies in number of breaths or seconds (less than 30 breaths or seconds) facilitate the identification of the plateau phenomenon without reducing the reliability of the measurements. When employing metabolic carts, averaging intervals including 15-20 breaths or seconds are preferable as a compromise between capturing the true O and identifying the plateau. In training studies, clinical interventions and meta-analysis, reporting of O in absolute values and inclusion of protocols and the averaging strategies arise as imperative to permit adequate comparisons. Newly developed correction equations can be used to normalize O to similar averaging strategies. A lack of improvement of O with training does not mean that the training program has elicited no adaptations, since peak cardiac output and mitochondrial oxidative capacity may be increased without changes in O.