That's not entirely accurate. On stock wheels the vehicle speed is 117 at an indicated 120 (tested on a buddy's car, same as mine, that's stock standard). On my car it's 116 at an indicated 120. Used the GPS reading to determine the actual speed.
Unfortunately there is no formula where "one size fits all" in order to calculate the inaccuracy.
When we were doing regularity rallies where your odometer is blanked out and all calculations and instructions are based on you doing dead accurate speeds, we found that we had to re-calibrate our speedos before each event, and then on the longer events like Joburg to Swaziland, the readings were already becoming inaccurate as we started nearing the end, simply due to tyre wear.
Originally we would test the speedo in 5 km/h increments from 10 km/h up to 120 km/h over a measured kilometre, and get 100% accurate readings that way...then on an event, we would use the 200 m boards on the main roads to double check speeds. We HAD to do it because you can be told to do 73 km/h for ten minutes and twenty five seconds, and then turn left....if there were two roads to the left, say within 100 m of each-other, you were totally buggered if you were actually doing 75 instead of 73!!
Anyway, I digress...best way to do it is to check with a GPS as that is probably more accurate than a stop watch and line of site on 200m boards. For every day driving, it is most certainly the best option, and a variation of 1 or 2 km/h in either direction shouldn't make any difference.
What we DID find on the rallies, was that there was absolutely no guarantee that you could take two Golfs, for example, with same size tyres, and expect them to both have the same speedo error.
So, that being said, don't look at Car Magazine road test and say, oh, I drive a BMW M3 and it has a speedo error of x%...the car that was tested had that error...not yours....
Hope that all makes sense.
