That would affect only a few borderline trials and would average out with subsequent throws. It would be much more worrisome that the length of a needle or the width of a stripe is not infinitely precise, that would consistently affect all the trials.