How the long distance runner plans to reach the finish line in the shortest possible time remains elusive as it is determined by complex interaction between psychological and physiological factors. Yet, the athlete is so accurate in expressing his capacity that A. V. Hill stated, ‘some of the most consistent physiological data are contained, not in books on physiology, not even on medicine, but in world's records for running different horizontal distances’ (Lloyd, 1966). A. V. Hill used these records to formulate a mathematical model with the energy required for running expressed as the sum of a store and a continuous provision. Physiological indices describe work capacity in terms of pulmonary oxygen uptake and accumulation of lactate in blood although the distributing volume and the elimination of lactate are not regularly evaluated. A quantitative measure of anaerobic capacity, albeit too large, is the part of the total energy demand that is not covered by the pulmonary oxygen uptake during exercise and expressed as the (maximal) oxygen deficit, while others find the workload that elicits a given blood lactate level (often 4 mm) to be a sensitive predictor of performance. Also the ability to excel over long distances relates to the muscle glycogen level and genetic variation comes into play, e.g. with the percentage of slow twitch muscle fibres being higher in the endurance athlete than in the general population (70–80%versus 50%). Furthermore, training enhances not only aerobic capacity but also the number of capillaries around each muscle fibre, and hence the oxygen diffusion into the cells is enhanced 2.5-fold. Such physiological descriptions of the athlete, however, provide little help in choosing the optimal race pace while the most obvious candidate for that purpose is the ventilatory response. Ventilation increases curvilinearly with work rate to impressive levels as it may approach 270 l min−1 during whole body exercise (Jensen et al. 2001). Several mathematical procedures have been used to define the work rate that elicits the progressive increase in ventilation, but for the athlete that work intensity simply manifests when he cannot talk without being dysphonic. The study by Amann et al. (2009) in this issue of The Journal of Physiology has taken up the challenge to evaluate whether neural influence from the exercising muscles is integrated in the central nervous system and contributes to the choice of work rate during a 5 km ergometer cycling time trial. The same group (Amann et al. 2008) found in a similar study increased motor output during exercise where the neural feedback from the working muscles was attenuated by epidural anaesthesia, but since the local anaesthetic also affected muscle strength, time trial performance was reduced. In contrast, intrathecal administration of the morphine analogue fentanyl did not affect muscle strength and performance was enhanced by 6% over the first half of the maximal effort but then decreased by 11% over the second half of the trial (Amann et al. 2009). Obviously, the athletes overestimated their work capacity at the beginning of the trial possibly because less information reached the central nervous system from the working muscles in response to the intrathecal administration of fentanyl. Yet, at least one other interpretation is relevant besides influence from, e.g. skin, body and notably brain temperature (Nybo et al. 2002). With the administration of fentanyl, the ventilatory response to the larger workload accomplished in the first half of the trial was reduced by almost 4%. Thus, if we accept that ventilation provides an important feedback for the choice of the work intensity, it may be that this attenuated signal indicated to the athlete that he could increase the pace. Before we concur with Amann et al. (2009) that afferent signals from the muscles are crucial for choosing the optimal work intensity, at least one more piece of information needs to be provided. It should be shown whether the same elevated initial work load is chosen while ventilation is maintained at the same level as in the control trial, e.g. by administration of doxapram. Yet, the unique work by Amann et al. (2009) opens an avenue for evaluating, in physiological terms, the experience of the athlete while previous studies describe performance in terms of the relative work load where heart rate, oxygen uptake and blood lactate are the most often used references.