During the concentric movement of the bench press, there is an initial high-power push after chest contact, immediately followed by a characteristic area of low power, the so-called "sticking region." During high-intensity lifting, a decline in power can result in a failed lift attempt. The purpose of this study was to determine the validity of an optical encoder to measure power and then employ this device to determine power changes during the initial acceleration and sticking region during fatiguing repeated bench press training. Twelve subjects performed a free weight bench press, a Smith Machine back squat, and a Smith Machine 40-kg bench press throw for power validation measures. All barbell movements were simultaneously monitored using cinematography and an optical encoder. Eccentric and concentric mean and peak power were calculated using time and position data derived from each method. Validity of power measures between the video (criterion) and optical encoder scores were evaluated by standard error of the estimate (SEE) and coefficient of variation (CV). Seven subjects then performed 4 sets of 6 free weight bench press repetitions progressively increasing from 85 to 95% of their 6 repetition maximum, with each repetition continually monitored by an optical encoder. The SEE for power ranged from 3.6 to 14.4 W (CV, 1.0-3.0%; correlation, 0.97-1.00). During the free weight bench press training, peak power declined by approximately 55% (p < 0.01) during the initial acceleration phase of the final 2 repetitions of the final set. Although decreases in power of the sticking point were significant (p < 0.01), as early as repetition 5 (-40%) they reached critically low levels in the final 2 repetitions (>-95%). In conclusion, the optical encoder provided valid measures of kinetics during free weight resistance training movements. The decline in power during the initial acceleration phase appears a factor in a failed lift attempt at the sticking point.