I'm going to agree with the general sentiment expressed here, that the data display from a simple bike computer will have inherent inaccuracies. And this is okay. One can calibrate out errors from the wheel's circumference under load (roll-out 3x and then average isn't a bad method). But what are the other sources? Sampling rate error from the sensor is most likely the next largest. Next will be rounding/truncation errors in the algorithm's math stack. The filtering algorithms will contribute negligible errors. Last will be display resolution. Does any of this matter to the average bicycle rider?
GPS-based systems have their own sources of error. The math's more complicated, too. Does this matter?
It's not the mission to Mars. It's a bike ride.