Background: The aim of this study was to examine whether the addition of very light loads for modeling the force–velocity (F–V) relationship during the bench press (BP) exercise can confirm its experimental linearity as well as to increase the reliability and concurrent validity of the F–V relationship parameters (maximum force (F0), maximum velocity (V0), F–V slope, and maximum power (Pmax)).
Method: The F–V relationship of 19 healthy men were determined using three different methods: (I) 6-loads free method: six loads performed during the traditional free-weight BP exercise (≈ 1–8–29–39–49–59 kg), (II) 4-loads free method: four loads performed during the traditional free-weight BP exercise (≈ 29–39–49–59 kg), and (III) 4-loads Smith method: four loads performed during the ballistic bench press throw exercise in a Smith machine (≈ 29–39–49–59 kg).
Results: The linearity of the F–V relationship was very high and comparable for the three F–V methods (p = 0.204; median Pearson’s correlation coefficient (r) = 0.99). The three methods were ranked from the most to the least reliable as follows: 6-loads free (coefficient of variation (CV) range = 3.6–6.7%) > 4-loads Smith (CV range = 4.6–12.4%) > 4-loads free (CV range = 3.8–14.5%). The higher reliability of the 6-loads free method was especially pronounced for F–V slope (CVratio ≥ 1.85) and V0 (CVratio ≥ 1.49) parameters, while the lowest difference in reliability was observed for F0 (CVratio ≤ 1.27). The 6-loads free and 4-loads free methods showed a very high concurrent validity respect to the 4-loads Smith method for F0 and Pmax (r ≥ 0.89), a moderate validity for the F–V slope (r = 0.66–0.82), and a low validity for V0 (r ≤ 0.37).
Discussion: The routine testing of the F–V relationship of upper-body muscles through the BP exercise should include trials with very light loading conditions to enhance the reliability of the F–V relationship.