So, I would like to know why the magic is behind using 4:1 and 19 AWG?
A parallel elsie (LC) circuit has a high impedance at the resonant frequency, while your receiver has a low impedance. This leads to an impedance mismatch and poor signal power transfer from the elsie circuit to the receiver.
A 4:1 transformer performs impedance matching, which results in the receiver getting more power from the elsie circuit.
Using thicker wire for the elsie inductor coil reduces heating power losses due to wire resistance. The elsie circuit requires low-resistance wire to achieve a higher Q-factor. A higher Q-factor provides a higher noise-free amplitude gain but reduces the bandwidth.
An ideal parallel elsie circuit would have infinite impedance, infinite Q, and zero bandwidth. However, since no wire has zero resistance in real world, you can't achieve infinite Q, so amplitude gain is reduced, and the bandwidth is more than 0 Hz. The optimal bandwidth for an AM receiver is about 20 kHz, but tuning the antenna for each frequency is required, so there is a compromise between antenna bandwidth and Q-factor.
A full-size half-wave dipole antenna typically has a Q-factor of around 10-20. High-Q magnetic loop antennas can have a Q-factor of up to 2000, which, for a carrier frequency of 7 MHz, corresponds to a bandwidth of 3.5 kHz, which is sufficient for a voice reception.
Q = 1 / BW
BW = BW_Hz / fc
BW - relative bandwidth
BW_Hz - absolute bandwidth in Hz
fc - carrier frequency in Hz
combine it we get:
Q = fc / BW_Hz
For example:
1) BW_Hz = 3.5 kHz, fc = 7000 kHz:
Q = 7000 / 3.5 = 2000
2) BW_Hz = 500 kHz, fc = 7000 kHz:
Q = 7000 / 500 = 14
As you can see, this is why a usual half wave dipole at 7 MHz has bandwidth about 500 kHz and Q about 14 and small magnetic loop antenna at the same 7 MHz has bandwidth just 3.5 kHz and Q about 2000.