To test the hypothesis that 'live high-base train high-interval train low' (HiHiLo) altitude training, compared to 'live low-train high' (LoHi), yields greater benefits on performance and physiological adaptations. Sixteen young male middle-distance runners (age, 17.0 ± 1.5 y; body mass, 58.8 ± 4.9 kg; body height, 176.3 ± 4.3 cm; training years, 3-5 y; training distance per week, 30-60 km.wk-1) with a peak oxygen uptake averaging ~65 ml.min-1.kg-1 trained in a normobaric hypoxia chamber (simulated altitude of ~2,500 m, monitored by heart rate ~170 bpm; thrice weekly) for 3 weeks. During this period, the HiHiLo group (n = 8) stayed in normobaric hypoxia (at ~2,800 m; 10 h.day-1), while the LoHi group (n = 8) resided near sea level. Before and immediately after the intervention, peak oxygen uptake and exercise-induced arterial hypoxemia responses (incremental cycle test) as well as running performance and time-domain heart rate variability (5-km time trial) were assessed. Hematological variables were monitored at baseline and on days 1, 7, 14 and 21 during the intervention. Peak oxygen uptake and running performance did not differ before and after the intervention in either group (all P > 0.05). Exercise-induced arterial hypoxemia responses, measured both at submaximal (240 W) and maximal loads during the incremental test, and log-transformed root mean square of successive R-R intervals during the 4-min post-run recovery period, did not change (all P > 0.05). Hematocrit, mean reticulocyte absolute count and reticulocyte percentage increased above baseline levels on day 21 of the intervention (all P < 0.001), irrespective of group. Well-trained runners undertaking base training at moderate simulated altitude for 3 weeks, with or without hypoxic residence, showed no performance improvement, also with unchanged time-domain heart rate variability and exercise-induced arterial hypoxemia responses.