Do remember though I am refering to wear and mating cycles, these figure reflect that and I can believe them.
I can't say that a big USB would get used to the extent of 1500 mating cycles so how can you compare.
This isn't a port design issue, I feel the port is a lot better design; for the reasons above.

The problem as you say is the way the port has been surface mounted to PCB's.
I do agree that these tests must have only included the port internals and not the mounting to the device.
I feel that the methods behind mounting the USB port was not throughly tested before being forced on to the market as a compulsary requirement; so the most cost effective was used.