Well, let's think through that a bit, using some critical thinking skills. When the guy measures the marks, its 2,155mm apart, or 84.84". Divide that by pi (3.14159), and you get a rolling diameter of 27.01". Now divide that by 2 to get the rolling radius - 13.51". After he deflates the tire and repeats it, there may be a slight difference. It's very hard to tell exactly how close it is from that video, but I think most people would say that difference is fractions of an inch at most. Let's be conservative and say it's 1/2" shorter, so 84.34". Divide that again by pi and you get a rolling diameter of 26.85". Now divide that by 2 to get the rolling radius - 13.42". So, on the radius, that's a difference of 0.09". If the tire on the left has a rolling radius of 13.51", does the tire on the right look like it's got a rolling radius of 13.42"?
View attachment 349685
Scaling those photos, if the tire on the right has a rolling radius of 13.51", the one on the right has a rolling radius of about 12.20". Now, let's go backwards. Take that rolling radius and multiply by 2 and then by pi to get the rolling circumference - 76.65", which is 8.19" shorter than the first measurement, and there's no way that the "error" you point out is even of that order of magnitude.
So, what is the conclusion from the video? Rolling radius (as shown in the picture above) isn't really related to the rolling circumference (the distance between the marks), and the reason for this is that the tire's circumference is fixed and doesn't appreciably change with inflation pressure.
I'm still going to do Mr. Bills' experiment this weekend, and I am confident I'll get similar results.
Oh, and I don't run drag tires on my TJ. Do you? DOT radials are designed not to expand, unlike drag tires. If tire rpm mattered, you wouldn't be able to calibrate a TJ speedometer by changing a gear in the speedometer drive. The speedometer would only be accurate at one particular speed.