As I say this just my view of things.
For an item that is not entrenched into the installation such as your 300mm section of copper pipe in a bathroom then testing against a known earth should demonstrate whether or not it is extraneous.
Admittedly I am not sure how you got those readings there must be some sort of other connection in place somewhere, especially as once you had provided limited supplementary bonding the readings changed.
Yes those readings are curious, I suspect push fit plastic under the floor in the bathroom, but can't understand how connection to the cold water service did not improve the readings, but cross bonding to the hot water service did. The owner is going to refurbish the bathroom, so I might find out then!
Where there is a main service, say, entering in plastic, changing to metal then potentially going underground again and then around the house; to test against a known earth in the house would be likely to give an unreliable reading due to cross connections of services and cpcs.
To test against a known earth that was not part of the installation should give a valid reading of connection to earth.
But then we are back to the question of deciding whether the pipe is extraneous or not, such like in the bathroom scenario above?
Since all earthed points in the installation should be connected together at the MET (and other points around the installation) then removing the main earth to test against (with the installation off) would readily provide an earth that was not connected to the installation. An earth rod outside the property would also provide an earth reference but would be less likely to be available.
It may well be that the supply earth is connected to the next house which has main bonding to the water supply and so this would give you another incorrect reading but only if the service between the houses was conductive in which case you would need to bond anyway.
I can see your point there.
Generally I think the published guidance will work for most cases, however in a few cases I think it will give a false impression of zero potential, you could then test for this false reading against an earth reference outside of the installation.