Purpose To analyze if live high–train low (LHTL) effectiveness is improved when daily training is guided by heart rate variability (HRV). Methods Twenty-four elite Nordic skiers took part in a 15-day LHTL study and were randomized into a HRV-guided training hypoxic group (H-HRV, n  = 9, sleeping in normobaric hypoxia, FiO 2  = 15.0%) and two predefined training groups sleeping either in hypoxia (H, n  = 9, FiO 2  = 15.0%) or normoxia (N, n  = 6). HRV and training loads (TL) were recorded daily. Prior (Pre), one (Post-1), and 21 days (Post-21) following LHTL, athletes performed a 10-km roller-ski test, and a treadmill test for determination of V ˙ O 2max was performed at Pre and Post-1. Results Some HRV parameters measured in supine position were different between H-HRV and H: low and high (HF) frequency power in absolute (ms 2 ) (16.0 ± 35.1 vs. 137.0 ± 54.9%, p  = 0.05) and normalized units (− 3.8 ± 10.1 vs. 53.0 ± 19.5%, p  = 0.02), HF(nu) (6.3 ± 6.8 vs. − 13.7 ± 8.0%, p  = 0.03) as well as heart rate (3.7 ± 6.3 vs. 12.3 ± 4.1%, p  = 0.008). At Post-1, V ˙ O 2max was improved in H-HRV and H (3.8 ± 3.1%; p  = 0.02 vs. 3.0 ± 4.4%; p  = 0.08) but not in N (0.9 ± 5.1%; p  = 0.7). Only H-HRV improved the roller-ski performance at Post-21 (− 2.7 ± 3.6%, p  = 0.05). Conclusion The daily individualization of TL reduced the decrease in autonomic nervous system parasympathetic activity commonly associated with LHTL. The improved performance and oxygen consumption in the two LHTL groups confirm the effectiveness of LHTL even in elite endurance athletes.