Not to bust yer chops - but somehow the first half of yer statement kinda nulls the second
It's commonly accepted that even the factory speedometer in a car has a comparatively large margin of error: Speedometer error, Wikipedia whereas even the basic modern GPS has only about a 0.1 knot (about 0.1 mph) variance if tracking four satellites: Global Positioning System, Wikipedia.
Throw in the problem that it's simply impossible to have a "correctly calibrated" bicycle speedometer with the fact that when discussing MBs we're often bickering over even 5mph increments - such speedometers are really only good for "ballparking" speed estimates at best. Even the "radar trap" signs aren't nearly as accurate as we've been lead to believe (target size and vibration have huge effects): Radar gun, Wikipedia, to the point where speeding tickets have been contested successfully with GPS data: GPS Data Used To Disprove Radar Gun In Speeding Trial
For just $10-20, sure a digital bicycle speedometer is a fun toy to give ya an idea of your speed - IF you can manage to get it set even vaguely correctly, but remember the faster you go - the larger the error, whereas the opposite is true with a GPS as it'll get more and more distinct tracking information beyond say a slow walking speed (about 3mph).
