The solution given in this case is the least-squares solution for both sires and herds. If we compare what was produced by this analysis with what obtained through BLUP we get a glimpse of how BLUP works. BLUP solutions takes into account the fact that sire information has less variability than the variance of lactation yield of a single sire’s daughter. Effectively the sires estimates are shrunk toward the mean (assumed 0). The amount of shrinkage is dependent on the amount of information available for the sire. For example, if we take the predictions for the 4 sires obtained by BLUP ( ) we notice that prediction for sire CC is better than the one for sires ZA and BB even if lactation yield for his daughters are the same (110) than those for sires ZA and BB. This is because we have more information available for that sire.