Actual and mature equivalent yields calculated by users may differ from those calculated by AIPL even when identical parameter files are used. The edits system used by AIPL is very complex, and records processed using that system have been checked and many errors corrected. It is not possible to recreate that system in BESTPRED, which assumes that user input is correct. For example, there is a limit placed on the amount by which a test day observation can differ from expectations based on preceding and succeeding observations. Such edits must be performed before the test day data are processed by BESTPRED, regardless of how you are using the program (as a standalone program or as a subroutine call in another program).
Herd average yields also affect the best prediction calculations. When Format 4 records are processed using the standalone BESTPRED program the herd averages are back-calculated from the 305-d mature equivalent yields and the cow's deviations, and are set to 0 otherwise.
See About this document... for information on suggesting changes.