What would be the best number of trials before making accurate predictions about an experiment?

a.1
b.20
c.5
d.10

20