Did you read my answer in one of the recent threads about the Jim Williams pulser? Do you still believe there is such a thing as 1 Mohm impedance at 1 GHz? Have you done the math to see what a cable with 1 Mohm characteristic impedance looks like so it can match this supposedly resistive 1 Mohm input impedance?
I read it, but I don't understand the whole impedance matching thing... power doesn't need to be maximized; just enough power needs to get through for the device to accurately and precisely measure the frequency responce and relative amplitude. Obviously adding a 1M resistor in series with the signal input will greatly attenuate the signal (potential divider), but the instrument can calculate what the amplitude would be if that resistor wasn't there and display that for the user instead... just like using a 10x probe instead of a 1x one.
At audio (or similar frequencies) in the past, power maximisation was what it was all about.
Modern audio line amplifiers have enough power gain that they can afford to play fast & loose with impedance matching,so often amps have o/p Z of around 20 Ohms or so,& no effort is made to match them to the next device,which may be 600Ohms input Z,or 5kOhm or greater,with very little in the way of problems.
At high frequencies,however,you can't get away with such tricks,as cable characteristics come into play.
With an incorrect termination,the cable,from being an essentially passive player,gets involved as part of the circuit determining the resulting signal characteristics you will see with your 'scope.
An incorrectly terminated length of cable may affect the amplitude & phase of various components of a complex signal,whereas a correctly terminated cable will show the original signal.
Probes are designed to work around this problem to some extent,so as to make everyday circuit tracing/troubleshooting possible.
There are,however,as others have said,severe limitations,because real 'scope input circuits cease to look like 1M Ohm,input Z,due to their inherent parallel capacitance.
Those 'scopes with UHF capablility are designed to use 50 Ohms input Z from the outset,but usually,if an internal 50 Ohms termination is available at all on the types of Oscilloscopes that us lesser beings use,it is available by operating a switch.