CBR954 hits the nail on the head: in general machining practice, having cases like these bolted together to a "torque" from a torque wrench produces not-real-even case compressions. After they are align-bored and released, the holes often grow during the second bolt-up, after the threads in the holes smooth out and the torque becomes different. This can easily generate 4 tenths of a thousandth difference.
I first discovered this on Honda cases circa 1974, when we align-bored at K0 sandcast with over 100k miles on it. The holes were a little bigger than the case numbers indicated (which you'd expect by then), but they were not in line, but almost half a thou. So, the cases were bored 0.010" and shimmed with pegged shims, and new Black bearings were installed after polishing the crank journals. The final Plastigaging came in around 0.0015" for the 5 bearings. The most remarkable thing happened, though: that engine was so smooth that we started calling it "the electric 750". It was by FAR the smoothest running 750 I have ever seen. If I had the $$, I'd do it to mine in a heartbeat!