I'm an undergraduate engineer who has taken stats and I've never even heard of MAD before. I've thought about it [mean(abs(x)) ] but I've always assumed it woul

I'm an undergraduate engineer who has taken stats and I've never even heard of MAD before. I've thought about it [mean(abs(x)) ] but I've always assumed it would behave similarly . Which also seems to be case (apparently MAD is better though I don't really see why, as I understood it his argument is practically the same as when arguing about using median vs mean), which is why I'm not really understanding what the problem is. It's just a rescaling, as long as one sticks with one measurement type what's the problem?