It's those dyslexic fingers agaoin....see?

Any GPS, whether differential or not is surely up to the relatively elementary job of average speed over, say, 1nm.
Even if the accuracy is altering it's not going to alter by enough over a couple of minutes to vastly alter the average figure is it? I always believed that GPS (non differential) was only ever 'out' by a max of 50metres or so and that the accuracy didn't wildly swing about with any rapidity.