The last two engines with ARP 2000 rod bolts (Molnar rods) had around 0.006 variance, so the gauge was different on nearly every bolt.

I feel like the wording is vague, and don't think Jeff Smith was actually suggesting measuring the bolts without loosening and measuring the change. The way I read the article, you take a baseline measurement, loosen the bolt, recheck the length to calculate change and that will tell you if they were torqued correctly or not. Odds are if 1-2 rod bolts are correct, you don't have to check all 16. I pasted the paragraph below for quick reference. He didn't say look up the length as much as look up the original stretch requirements... I actually challenge that approach from a different angle because ARP typically defers required rod bolt stretch to the rod manufacturers.

This technique can also be used to check the veracity of an engine that you may not know the history of. If the rod bolts are from ARP, you can look up the part number and then search for the online instructions on the ARP website to find the original stretch numbers and compare these numbers to what you see on the actual bolts. You may be surprised at what you find, since many engine builders do not use rod bolt stretch when torquing the rod bolts.