The reason I did what I did is because of my explanation below and why I don't see Yamaha's process any different.
I work in a manufacturing facility where we do exactly this: fasten two parts together and bore them to within tight tolerances.
For the sake of explaining, I'll relate our process at work with names of cylinder head parts.
We press the majority of the dowels surface into the cap so it retains in the cap for two reasons:
1. The cap has an open hole so it can be accessed from the outside (and, well duh, there's a fastener going through it). Hence, if my occurrence were to happen, they can pull the cap off, press out the dowel, and try again (which it does happen on seldom occasion). If you press the majority of the dowel into the cylinder head and something boogers up, how are you going to get it out? From an at-home perspective we can remedy something but in an assembly line, this can't be tolerated. I know it seems ironic that I ask how to fix my problem when we have these problems at work. Well, our dowels, caps, and cylinders heads are MUCH beefier so if a dowel is crushed, the other parts remain unscathed.
2. The cap is separated during another assembly process so if you leave the dowel in the cap, this alleviates damaging the dowels since people are working around the cylinder head. You're at risk if you have people assembling around thin walled dowels.
I'm not saying anyone is wrong but I can't understand the logic if the majority of the dowel is in the cap or the head (unless you're looking at it from a manufacturing assembly perspective). Flip the parts. Imagine, in a theoretical sense, mating the head to the caps; the caps being stationary and you're able to maneuver the head. It's not different.
It just eludes me because there wasn't any resistance; it fasted with 7 ft-lbs down to the mating surface. The dowels are also tapered on the ends.
Oh well, chalk one up for the weird cases of "when service maintenance goes wrong".