I don't know why that didn't jump out at me before. He's not measuring at the datum, but where the neck meets the shoulder. His numbers are about right for that.
I hate to beat a dead horse, since the OP reported the problem solved in post #16, but getting the basics of the measuring right seems worthwhile, and maybe this will help someone else trying to make the same measurement.
Mr. Guffey is correct that the normal measuring point is an intersect with the shoulder of a bottleneck cartridge called the datum line. It's actually the intersection of the outside of the shoulder with a plane through it and perpendicular to the axis of the case. This intersection describes a ring around the shoulder at 0.400" diameter. Here's a drawing I did awhile back to explain this:
The reason for using that datum is to get around having to estimate exactly where the vertex of the slight radius of a corner is. The old way of measuring headspace was where the slopes of the side and shoulder would intersect, but it requires that estimation factor and is a less repeatable measurement for that reason. Same with the neck and shoulder junction.
If you don't have a tool for measuring this exactly, comparative measurements approximate to the datum line are good enough for adjusting dies. You just won't necessarily get a number you can compare to anyone else's number. That's because this will hit the shoulder a little bit off from the datum line, so the lengths won't be exactly right. But if the hole through the spacer is anywhere close to the datum line diameter, then using this method to see how much you are pushing a shoulder back works just fine.
Below is an improvised measurement using a spacer from Lowe's that claims to be for 3/8" bolts. The actual ID is 0.409", IIRC, and that's close enough to the .400" datum diameter for comparative measurements of shoulder setback. You can see from the number, 1.594", when it should be about 1.640" that the spacer sets a little lower on the shoulder than an actual datum diameter hole would. This is mostly because of the chamfer in the hole, and not the fact it is a little large. Since you zero on the length of the spacer, that chamfer subtracts from the absolute value measurement.
Nick