on 15/02/2019, NY supposed :
I've always thought that timing over a fixed distance is subject to lots of
random errors such as the reaction time of the observer to press the button,
and his judgment as to the exact instant when he has passed the fixed
features such as the white squares painted on the road.

Back then, what better method could they use? A fraction of a second
late or early pressing the button, would not make that much difference
to the check over a 1/4 or 1/2 mile, but they used to allow a bit of
leeway anyway - making sure you were well over the limit.