When the factory assemblers are doing barrel/bolt fitting (and actually, if they're doing it by the manual, they're doing bolt group fitting) it's by a selective parts method:
You have a barrel torqued to the upper, you install a bolt group (minus ejector), slip a go gauge in the chamber and using thumb pressure ensure the end of the carrier is flush with the end of the upper. The go gauge is removed and a no-go gauge is inserted and the end of the carrier should stick out, very slightly, from the end of the upper. If it doesn't, try another carrier group after placing the failed carrier group in the minus tolerance bin.
It's a whole lot easier to sit there with a barrel, gauges and a basket of bolts (w/o ejector) and check fit. Again, minus bolts go in one bin, plus bolts another for trial on other barrels. Now, this is doing it right. It's a whole lot easier (and cheaper) for most of us to order a matched barrel & bolt.
I won't deny that in a great many cases, tolerances will be such that dumb luck can possibly give an acceptable fit. If you want to bet your gun and body parts on it, feel free. My question: have you ever actually gauge checked your fit? There's roughly a 0.003 in headspace difference between a go and no-go gauge. There's another roughly 0.003 in headspace difference between a no-go gauge and a field gauge. The military is happy if used barrels & bolts won't accept a field gauge. That's not the standard you want in a new assembly.
Note: our lead armorer serial numbers the BCG to the rifle. Headspace checks are made per vendor requirements. When bolt replacements are made, the combination is gauge checked.