Purpose <p>The short-term scaling exponent alpha 1 (DFA α1), derived from detrended fluctuation analysis of heart rate variability, quantifies self-similarity within interbeat interval time-series data. In exercise physiology, DFA α1 has primarily been studied during incremental exercise, but less is known about its application to training load. We examined the effect of exercise duration on DFA α1, relationships between DFA α1 and internal training load metrics, and the repeatability of DFA α1.</p> Methods <p>After performing a baseline ramp incremental test and time to task failure (TTF) trial, 15 (10M and 5F) healthy recreationally active participants performed 20-min and 40-min constant-work rate tests (CWR₂₀ and CWR₄₀), each followed 5&#xa0;min later by a TTF trial. DFA α1 was measured during each session and at a fixed, moderate intensity (MOD) before and after CWR and TTF trials.</p> Results <p>DFA α1 decreased significantly during CWR₄₀ (p &lt; 0.001) and was lower at the end of CWR₄₀ compared to CWR₂₀ (p &lt; 0.001). DFA α1 measured during MOD significantly decreased (p &lt; 0.001) after CWR₂₀ and CWR₄₀, with a larger decrease after CWR₄₀ (p = 0.018). The decrease in TTF relative to baseline (i.e., acute performance decrement) and training load metrics were greater for CWR₄₀ than CWR₂₀ (p &lt; 0.001); however, these metrics were not correlated with DFA α1. Nevertheless, DFA α1 was repeatable across days.</p> Conclusion <p>DFA α1 measured during or in proximity to a task was sensitive to exercise duration but unrelated to individual differences in internal training load, indicating further refinement is needed.</p>