When witness accounts completely contradict each other "cherry-picking" must occur.
No. Cherry-picking is never appropriate.
When witness accounts contradict each other, one still has to use a rational fact-finding process to determine what happened. If witness perception is skewed due to a common factor, such as sound reflections interfering with a witness' perception of the direction of the source of the sound, you may see large groups of witnesses disagreeing with each other. But if there are no common factors that would induce a common error in witness observations, then errors will tend to be random and accurate observations will agree with each other. That is just common sense and common experience. And we see this in the evidence in this case.
It is not cherry-picking to find that the 80% who said 3 shots over a matter of several seconds should be preferred to the recollection of Jean Hill who thought there were more than 3 and less than 7 shots, or to A.J. Millican who thought there were 8 shots over 5 minutes, or the handful of witnesses who could only recall two shots or thought there were 4 shots. The distribution of shots in the shot counting fits exactly with what one would expect if there were exactly 3 shots recalled correctly by the vast majority with conflicting witnesses making errors that are randomly distributed over the other counts.
It is not cherry-picking to conclude that the witnesses who said that the shots appeared to come from the TSBD are to be preferred over the witnesses who said they appeared to come from somewhere farther south. There are very sound, rational reasons for concluding that they all came from the SN and no corroborating evidence that any shots came from somewhere else. The confusion that is evident in the number of witnesses who thought the sound came from other directions is consistent with the sound reflections in Dealey Plaza interfering with a human's ability to determine sound direction.
It is not cherry-picking to conclude that the third shot and last shot struck JFK in the head. It is not cherry picking at all to observe that the preponderance of evidence on that issue favours the head shot being the last. While Charles Brehm thought there was a shot after the head shot, there is nothing to corroborate that and much conflicting. Altgens was adamant that the head shot was the last. The Secret Service Agents, the Connallys, the police officers riding close to the President all agreed. Emmett Hudson initially mentioned only a shot hitting JFK when the car was in front of him (he was on the steps going up the knoll). It was only in his WC testimony that he mentioned a shot after the headshot but his testimony was so at odds with his earlier statements that it is difficult to place much reliance on his WC testimony at all.
It is not cherry-picking to conclude that the second shot struck Governor Connally. There is consistent corroborative evidence from those closest to the events (the accounts of the Connallys, bystander Gayle Newman, and the Secret Service agents and police officers around and behind the President) and no conflicting evidence that I have found.