Purpose: To explore the effect of several methodological factors on the number of repetitions performed before and after reaching certain velocity loss thresholds (VLTs). Method: Fifteen resistance-trained men (bench press 1-repetition maximum = 1.25 [0.16] kg·kg⁻¹) performed with maximum intent a total of 182 sets (77 short sets [≤12 repetitions] and 105 long sets [>12 repetitions]) leading to failure during the Smith machine bench press exercise. Fifteen percent, 30%, and 45% VLTs were calculated, considering 2 reference repetitions (first and fastest repetitions) and 2 velocity variables (mean velocity [MV] and peak velocity [PV]). Results: The number of repetitions performed before reaching all VLTs were affected by the reference repetition and velocity variable (P ≤ .001). The fastest MV and PV during the short sets (75.3%) and PV during the long sets (72.4%) were predominantly observed during the first repetition, while the fastest MV during long sets was almost equally distributed between the first (37.1%) and second repetition (40.0%). Failure occurred before reaching the VLTs more frequently using PV (4, 8, and 33 occasions for 15%, 30%, and 45% VLTs, respectively) than MV (only 1 occasion for the 45% VLT). The participants rarely produced a velocity output above a VLT once this threshold was exceeded for the first time (≈10% and 30% of occasions during the short and long sets, respectively). Conclusions: The reference repetition and velocity variable are important factors to consider when implementing VLTs during resistance training. The fastest repetition (instead of the first repetition) and MV (instead of PV) are recommended.