This shows how accurate the normal approximation to the Binomial is for various values of the success probability with 30 trials.
Try moving the sliders for t and p to see how they impact the accuracy of the normal approximation. Does the continuity correction seem to work?