I totally agree that the manufacturer (the one who did the testing and who has to maintain quality control of their product through periodic testing among many other duties) is going to be able to provide the best advice for how their products are best installed and maintained. Their best torque value will be based on either lubricated or dry assembly, i.e. what torque value worked best with that particular method of assembly.
I've known certain brands of cranks that were susceptible to cracking at the left crankarm square taper area, and it became known that they tended not to crack if torqued only once instead of periodically.
The bolt tension that exists immediately after torqueing the bolt is not the same value as after the bottom bracket has exerted torsion on the crankarm through the taper on certain cranks that simply cannot sustain quite that level of bolt tension over time without creeping up the spindle tapers. Some Shimano cranks with visibly more metal in this area may actually sustain the full tension value of the freshly-torqued bolt in use over time, but some do not, as can be felt by the bolt turning further each time that the bolt is torqued. Some cranks seem able to take it, and some (usually the lighter ones with less metal surrounding the spindle) can not.
Even some Shimano cranks that came after the earlier Deore cranks of the late 80's and early 90's would tend to walk up the spindle, and at one point Shimano began installing steel stop rings on their bottom brackets so that at least the drive side crankarm would not move significantly over time relative to the front derailer. And there is another issue having to do with how much twist is occurring along the tapered ends of the spindle inside of the hole in the crankarms, something that is much exacerbated by riding over rougher surfaces or jumping the bike with the cranks horizontal, as this definitely has the effect of relaxing the bolt's tension as more flexing movement occurs between the spindle and crankarm.