I have no doubt that your connection to the chassis at the rear of the ML was probably done better than I would have done and I am positive that the chassis is more than capable of carrying the 50 or so amps that your compressor would have needed. In fact it would handle a couple of thousand amps before it started melting, but my point is it was never designed to carry that current.
What did you do at the front end of the chassis? Or did you rely on all the metal to metal connection points joining the chassis to the body to the motor to the battery. All the major structural connection points between chassis and body are isolated, engine mounts gearbox mounts so little or no current is going to flow through then. Body mounts are bolted through but are subject to dirt oxidation etc i would not rely on them etc etc.
I have no doubt here is probably some point of electrical connection between body and chassis, but it will not be big enough to handle the starting current of your compressor, or in my case the starting current of the engine if I parallel up my batteries.
To design the chassis to carry current you need to do something at both ends, and in most cases it is better, easier, and less work to run a cable back to the negative post on the battery.
A lot of people, as you know, think putting a cable under a bolt head is sufficient and away we go. And this is where I come in. Mates 4wd’s where the beer is not cold when they go away, Excavators, Trucks, plant equipment where the eutectic fridge or UHF or GPS or range finder or depth measurement or weigh system is not working is, most times, traced back to fault 0V connections. I have seen some shocking installs
The monocoque nature of the Pajero is a different situation altogether and makes most of this argument moot