Our goal was to quantify the change in BW (ΔBW) per change in body condition score (ΔBCS). Our database had 69 studies of 2,181 Holstein cows (75 to 263 DIM) with weekly records of BW, BCS, and DMI. The average interval between consecutive BCS for each animal was 38 d. Means and standard deviations were 644 ± 83 kg BW, 3.1 ± 0.4 BCS, and 25 ± 4 kg DMI with 12 ± 22 ΔBW, 0.07 ± 0.18 ΔBCS, and 0.0 ± 2.6 ΔDMI for the period. Using HPMIXED from SAS 9.4, a mixed regression with random effects of animal, study, and location and fixed effects (as described next) was fitted to determine the ΔBW per ΔBCS. To determine the best model to predict ΔBW, we developed 4 candidate models with 2 levels of complexity for fixed effects with and without effects of parity. A simple model contained only linear effects of initial BCS, initial BW, ΔDMI, and ΔBCS, whereas a complex model also included 2-way interactions of all main effects and quadratic and cubic polynomials. Cross-validation across study was used to select the best fitting model based on concordance correlation coefficient (CCC), root mean square error of prediction (RMSEP). Parity was highly significant but had little impact on CCC and RMSEP. Model complexity also had little impact on CCC but the simple models had lowest RMSEP. The interaction of ΔBCS with initial BCS was not significant (P = 0.8), so we conclude that ΔBW per ΔBCS does not depend on starting BCS. Our proposed model is ΔBW = −17 + 0.042 × BW + 2.7 × ΔDMI + 54 × ΔBCS + PARITY_i (CCC = 0.77, RMSEP = 20), where PARITY_i of 1, 2, and 3 or more, adds 0, −3.6, and −5.9, respectively. With this equation, a 1-unit ΔBCS is associated with ΔBW of 54 kg, which equals 8.4% of live BW or 9% of empty BW (assuming gut fill is 5.2 x DMI). Using the same statistical approach for early lactation (<75 d-in-milk), ΔBW is 64 kg per 1-unit ΔBCS, or 10% of live BW. This is substantially less than the value used in the NRC 2001 of 14%. These data should be limited to the typical ΔBCS of most cows, which range from BCS 2.0 to 4.0.