I am a new shooter, but it would seem to me that all classifers are "out of wack" at some level. The question is not, is it out of wack, but is it out of wack enough to make a difference in classification levels, does it make a B shooter a C or an A. If you really wanted to do it right you could look for significant changes in class throughout the day and see if they clustered in the times in question. You probably wouldn't have enough samples to get any true statistical power but you could look for a trend. Voting does no good since everyone who did well votes yes those that did poorly no. It seems like the MD should be figuring out did this deviation rise to the level where aberant classifications will result.