use normal approximation to binomial distribution and apply continuity factor (+ or - 0.5) because approximating discrete distribution (binomial) with a continuous one (normal).
z = (x - np)/(sqrt(npq), where = 102, p = 0.5 and q = 0.5 and x = number defective:
z = (x - 51)/5.05
P(x = 47) = P(46.5 < x < 47.5) = P( (46.5 - 51)/5.05 < z < (47.5 - 51)/5.05) = (-0.89 < z < -0.69) = P(z < -0.69) - P(z < -0.89) = 0.2451 - 0.1867 = 0.0674
P(x < 47) = P(x < 46.5) = P(z < -0.89) = 0.1867
P(x > 47) = 1 - P(x = 47) - P(x < 47) = 1 - 0.1867 - 0.0674 = 0.746
P(x = 57) = P(56.5 < x < 57.5) = P((56.5 - 51)/5.05 < z < (57.5 - 51)/5.05) = P(1.09 < z < 1.29) = P(z < 1.29) - P(z < 1.09) = 0.9015 - 0.8621 = 0.0394
P(x >= 57) = P(x > 56.5) = P(z > (56.5 - 51)/5.05) = P(z > 1.09) = 1 - P(z < 1.09) = 1 - 0.8621 = 0.1379
P(x <= 57) = P(x < 57.5) = P(z < 1.29) = 0.9015