Dear Editor-in-Chief, In their letter, Marwood et al. conclude that critical power (CP) rather than maximal lactate steady-state (MLSS) offers the best representation of the maximal metabolic steady-state (i.e., critical intensity). While defending this position, the authors claim: That we previously proposed (1) “CP…overestimates the heavy–severe boundary.” This is a distortion of our views. Although overestimations predominated in Mattioni Maturana et al. (2), we contend that CP can both overestimate and underestimate the maximal metabolic steady-state depending on factors, such as the type of test, model, and fitting strategy used (3–5). In fact, an approximately 5% margin of error in CP estimation is acknowledged by others who support CP as the heavy–severe intensity boundary (6). While acknowledging that MLSS is itself an estimate of the maximal metabolic steady-state, in our view, MLSS is superior to CP testing because it simultaneously verifies whether the physiological responses conform to those expected at the critical intensity. Using MLSS “as the primary marker of the heavy–severe intensity boundary…is ironic given the arbitrary and highly liberal definition of MLSS.” There is no irony in using a delta change of 1 mmol·L−1 between 10 and 30 min as the criteria for a stable [La]. This is a well-established model. How liberal or conservative this measure needs to be can be debated, but normal measurement variability must be considered. Nevertheless, providing a physiological validation of the critical intensity of exercise is always more appropriate than accepting a model parameter estimate without any verification. “Based on available evidence…we contend that CP, when appropriately determined, is most representative of the upper limit of the metabolic steady state” This observation simply ignores several recent lines of evidence (2,3,7) and even common sense (i.e., how can metabolic steady-state be assumed without measuring metabolic responses to exercise?). We have discussed this topic in detail elsewhere (8). In short, Poole et al. (6) defined CP as “the highest intensity that can be sustained for a prolonged time solely by oxidative energy provision.” In this definition, exercise at CP does not draw upon anaerobic metabolism. Thus, progressive depletions in phosphocreatine and accumulations of [La] are not evident with time, which minimizes metabolic and acid–base disturbance and delays the initiation of fatigue. Therefore, the physiological responses expected at CP are those of MLSS. Any differences between these two indices simply relate to imprecisions inherent with the methods used for their determination. Although we believe that CP is a good approximation of the heavy–severe boundary, the clear limitations of this approach (2–4) make its use for research purposes inadequate, unless physiological validation is conducted to confirm metabolic stability at CP. To conclude, too often it is assumed that the model output estimate of CP reflects the true critical intensity of exercise (i.e., the heavy–severe boundary) despite compelling evidence that this is not always the case (2,3,7). Both MLSS and CP testing have inherent limitations, but in the absence of physiological verification, CP testing carries a greater degree of predictive uncertainty.