It just seems weird that so much attention to details was paid to reactor construction, etc., then the output result left to chance to a large degree by using a flimsy paper tube as the “calibrated” orifice from which all output measurements were taken.
It is not flimsy. It is stiff, and well formed. Lots of things in Japan are made of paper. Heck, in old houses the walls are made of paper, which is why your toothpaste is frozen the morning. * Nothing is left to chance. It has been carefully tested in traverse tests, which have been repeated many times, and we know it is right because the calibrations produce a balance of zero, after accounting for losses which are linear and unchanging.
* Okay, they also have sliding screen doors, and sliding wooden slats, deployed for snow and typhoons. In an old house, there are large gaps between the slats. Basically, living in an old Japanese house is like camping. It is picturesque but you wouldn't want to live that way, and there aren't many old houses left.