Interesting in a couple of respects:
1. He says “almost five seconds” “maybe four seconds” elapsed between the first and second shots.
2. He describes his reaction to the first shot as thinking someone had set off a firecracker and was scanning the crowd looking for such a person. He does not say he felt anything strike him. (He testified to the WC that he felt something strike his cheek on the second shot.)
4 - 5 seconds fits with a first shot at z180 to z198 and second at z271 leaving 2.3 seconds to aim and fire the last shot at z312-313.
You and I interpret Tague’s testimony differently wrt the first shot’s z-frame. This is a good example where not only can witness testimonies differ, but researchers can exacerbate the situation by interpreting any given testimony in a different way.
My effort to mitigate this: I understand that witness testimony can be quite variable. Witness variability was a problem I ran into early on when looking into the shooting timeline using ear witness testimony regarding how they all recalled hearing the first shot. To avoid this issue the PRT analysis was done without using any testimony, but rather based on human reactions, and the results came back saying a shot was triggered about z124. Someone I know challenged me by saying that even if witnesses have large variability in what was recalled, there still should be a subset of witness testimony that agreed with a first shot around z124.
The only way around this conundrum that I could think of was to put a tighter constraint on the witness testimony used, but not in a way that would bias any particular answer. This was done by using three shot witness testimonies that included both audible (hearing the first shot) and visual (positioned the presidential limo at the time of hearing that first shot with a fixed background landmark or photo view/camera reaction). This significantly reduces the sampling population but the thought was that testimonies “anchored” using 2 separate senses (sight and sound) would be more reliable with reduced variability.
So for that challenge, the technique tried consisted of (1) Using the “anchored” visual testimonies to estimate the limo’s position at the time of the first shot, and then augment that by (2) Using the power of averaging on those testimonies to help converge on a final estimate for the limo’s position on Elm at the time of a the first shot.
The results of this approach was very interesting. At the time of the first shot, the average limo location appeared to be around a z133 timing position (if using Phil Willis and his slide #5) or alternately slightly before a z133 timing position, right around z124, (if using Phil Willis and his #4 slide).
The approach was summarized at this link. I would recommend skipping over the background and
just click on the link that appears at the bottom of that page to see how the results converged. This approach used only a stricter testimony pool without any reaction time science.
https://sites.google.com/view/anchored-first-shot-testimony/home