There's a small positive bias in the extended source magnification calculation

[fbartolic]

There's a slight positive bias in the extended source magnification (less than $10^{-4}$ in magnitude):

The plot is error relative to VBB but the hexadecapole approximation confirms that the bias is real.