Which de-rating rule? And how can "proper design" overturn physics?
Insulation is damaged by heat. The types of insulation are rated in terms of what temperature they are prooven to withstand, for extended periods, without damage. These rating are the basis fo the tables, such as 310-16.
Electricity, encountering resistance within the wire, generates heat. This, along with an assumed 'normal' room temperature, also provide a basis for 310-16.
Now, what if the temperature where the conduit is is not "normal?" We de-rate, or sometimes up-rate, accordingly.
What if the load is only momentary? We use a larger ampacity requirement if the use is "continuous."
Now, what about the number of conductors? Another fact about alternating current is that the current flowing in one wire afects the current flow in another wire. We call this "reactive inductance" or "impedance." Mike Holt is said to demonstrate this by "shorting" both ends of 500 ft. of #14 wire on a 20 amp circuit. Rather than draw the 80+ amps simple resistance calculations would lead yo to expect, he measures onlt 12 amps. Why? His 500 ft. of wire is still wound on the spool.
This is also apparent with multi-wire circuits. For example, two out-of-phase hot wires, drawing 10 amps each, and sharing a neutral, will have nothing flowing in the neutral. Effectively, there are only 20 amps going through that pipe. Run a separate neutral with each hot, and you have ten amps through each of four wires- with a heat-generating effect of 40 amps running inside that pipe (4 wires at 10 amps each).
Often, we inadvertantly oversize our wires and our pipes. The NEC is nothing if not conservative. In our calculations, we assume current based upon our breakers (usually). That almost every circuit in a house will never see anything near 20 amps of load, electricians often get away with having huge bundles of romex, and not de-rating. They're not following code, but they get away with it.