Re: Why are physical checks needed?
The user experiences less than advertised connection and (rightly so) doesn't give the least about the cause of it, it is just the operator not delivering.
But it needs to be the result of *protracted* testing.
If a 90 second test at 14:15:09 last Thursday gave a result of 70Mbps download speed, would you be quite happy even though most of the time you hardly ever get more than 1Mbps?
While real-World tests are likely to be more accurate than the theoretical *if* they are done properly, the theoretical average performance is likely to be more accurate than a one-off 2 minute drive-through test.
Both data contention and signal path analysis are both very well established arts and give very accurate results *if* the correct parameters are used. Results that are more accurate than one-off snapshots of under 1% of the total area covered, anyway.