The regulation for speedometers (on mass-produced cars, at least) demands that they should never under-read (which is fair enough, I guess). They're allowed to over-read by up to 10% + 2.5 MPH. The result of that is that all car speedometers over-read slightly - typically 5% or so, in a modern car. The same regulation specifies odometer accuracy - which is +/- 4%, tested over a 7 mile run on a test track.
Trip computers, (in my experience at least), do tend to make the car look better than it really is. Mrs. Avocet's car is 2017 and the long term average fuel consumption is about 3MPG better than filling it brim-to-brim and recording it on a mobile phone app. If it says 43MPG, it has probably done about 40. Remember that we consumers willingly played along with the car industry "gaming" the system, because the same tests gave us the official CO2 figures that our tax was based on, so we paid less tax if the combined CO2 figure was lower than it really was! After the end of 2020, fuel consumption monitoring was regulated, so I imagine it's better on new cars today.
EV energy consumption monitoring is still in its infancy and the official tests need to be honed to make them more like real life. Why do you feel the need to add in charger / inverter losses for your electric car, but you're happy to only start counting fuel consumption for your ICE car once the fuel is in your tank? Why not add the transportation and pumping losses for your fuel between the refinery and your tank?
