Sorry, the Electrical Engineer in me feels compelled to respond:
Wire is more dependent on quality than size. Twisting is also anecdotal, it does not work in every case to help. I have run 60 feet of an n scale layout using 16 gauge high quality wire without twisting and have no issues. If you think about it, and 8 Amp controller with 15V track power is 120 WATTS or the old fashioned light bulb. And lamps use 16 gauge wire because it is 60HZ AC power. But they do not worry about a signal quality just power. And the wire is cheap because lamp cords are not very long and not good for the digital signals of DCC but they do not need to be.
The twisting adds a bit of capacitive reactance to possibly counteract the normal inductive reactance of the wire. This is thought to keep the waveform from degrading. The proper term is Dispersion. The fact is that most of the real source of dispersion is due to the impurities in the wire. And the larger the wire, the more impurities and the more reactance. There is a grave mistake using the AWG guidelines because they were empirically derived for DC and 60HZ AC, where the DCC signal is a composite of the 8K pulse repetition square wave constructed to contain modulated higher frequencies within the square waves. These other frequencies are much higher than the 8 K pulse frequency. This is why the AMG is not correct.
The fact that DCC has a limit of around 300 feet is because of the harmonics of the modulated information in conjunction with the 8K pulse. There is what is called a harmonic extinction function that is behind the length limitation. Near that point, the waveform becomes unrecognizable because it is getting close to maximum dispersion. This means the information in the signal is lost. These effects are not seen in most cases because the layouts are not close to that large.
PhD Electrical Engineering and Mathematics
Wire quality is more important than the size. And with high quality wire, twisting has no effect on signal dispersion. That is all controlled by the actual signal and its harmonics.