Oscilloscopes and the -3dB thing

[JustAnotherGuy999]

I understood that: if i have an oscilloscope with a 100MHz bandwidth and tried to visualize a waveform of 100MHz, 1V peak to peak, the oscilloscope is going to show me 70% of its real amplitude, i,e: 0.7V, is this true? if so is this mean that oscilloscopes are not to be "trusted" when it comes to amplitudes, and used just to see the waveform shape and frequency?
what is the maximum bandwidth that is considered to be "trustfull" in this oscilloscope?
is there a forumla how to calculate this bandwidth?
is there a forumla to calculate by how much the amplitude will diminish at a certain frequency?
if so why not the oscilloscopes makers "implement" it in the oscilloscope, and make life easier to every one?

[wraper]

-3dB is the worst case they can clam the bandwidth, usually they are much better.

[awallin]

If all you are measuring is sine-waves, then yes in principle you could correct for the amplitude roll-off and measure with better accuracy a bit beyond the specified bandwidth.
Any other waveform will contain multiple tones and the scope can't know what the signal contains so it's not feasible to apply corrections.

I've posted the Rigol DS2072A response in the upgrade thread, it deviates quite a bit from a simple RC-filter, so it's not something you can rely on -  you'd have to measure it for your particular scope.
http://www.anderswallin.net/wp-content/uploads/2015/12/rigol_DS2072A_upgrade.png