To quantify the accuracy of 5 equations to predict the metabolic cost of load carriage under ecologically valid military speed and load combinations. Thirty-nine male serving infantry soldiers completed thirteen 20-minute bouts of overground load carriage comprising 2 speeds (2.5 and 4.8 km·h21) and 6 carried equipment load combinations (25, 30, 40, 50, 60, and 70 kg), with 22 also completing a bout at 5.5 km·h21 carrying 40 kg. For each speed-load combination, the metabolic cost was measured using the Douglas bag technique and compared with the metabolic cost predicted from 5 equations; Givoni and Goldman, 1971 (GG), Pandolf et al. 1997 (PAN), Santee et al. 2001 (SAN), American College of Sports Medicine 2013 (ACSM), and the Minimum-Mechanics Model (MMM) by Ludlow and Weyand, 2017. Comparisons between measured and predicted metabolic cost were made using repeated-measures analysis of variance and limits of agreement. All predictive equations, except for PAN, underpredicted the metabolic cost for all speed-load combinations (p , 0.001). The PAN equation accurately predicted metabolic cost for 40 and 50 kg at 4.8 km·h21 (p.0.05), underpredicted metabolic cost for all 2.5 km·h21 speed-load combinations as well as 25 and 30 kg at 4.8 km·h21, and overpredicted metabolic cost for 60 and 70 kg at 4.8 km·h21 (p ,0.001). Most equations (GG, SAN, ACSM, and MMM) underpredicted metabolic cost while one (PAN) accurately predicted at moderate loads and speeds, but overpredicted or underpredicted at other speed-load combinations. Our findings indicate that caution should be applied when
using these predictive equations to model military load carriage tasks.