I think I see what you are trying to do, and I agree with you. But I had never thought my way through that exception before, so let me explain it in steps. Tell me if this matches your reasoning.
? You have three circuits, and the ampacities will be the same for all three.
? For any of the circuits, you have a five foot section that runs in a conduit that has a total of nine current-carrying conductors. You calculate an ampacity for that section, using the derating factors of table 310.15(B)(2)(a).
? YES, you do have to derate the conductors for that portion of the run. But wait, there is more!
? The remainder of the circuit is a 70 feet run that does not get derated, since there are no more than three CCCs in a cable or conduit.
? So the same circuit has different ampacities on one side or the other of the JB. The general rule of 310.15(A)(2) is that you have to use the lower of the two values. But wait, there is still more!
? The section with the smaller ampacity is less than 10 feet, and is less than 10% of the 70 foot (not the 75 foot, by the way) length of the section with the higher ampacity. The exception says you can use the higher value for the five foot distance from the transition point to the beginning of the run.
? Conclusion: You do have to derate, but you don’t have to use the derated (i.e., smaller) value of ampacity. This is a rare instance in which the same circuit has two separate ampacities for adjacent sections, and you get the use the higher of the two ampacities for the entire run.