Far too early to math all the numbers, but one note to this is that it's entirely possible to start with a median metric and switch to mean from there, their argument being how that's not possible right off the bat.
Start off at a median number, which statisticians love because it flattens out the effects of excessive weights on any particular segment, such as we have in the USA with a small number of obscenely rich people.
But then strip the anomaly from the dataset and recalculate the mean and get a proper average. The median may not move any meaningful amount, but without the anomaly the mean becomes more true to the sample.