Incorrect return value for EstimateFalsePositiveRate

[erikdubbelboer]

At line 77 EstimateFalsePositiveRate returns the number of false positives found divided by 100. This makes no sense as the number of tests performed depends on the input number. Choosing a higher n will now increase the number of false positives by a too high amount.

The correct return value would be the number of false positives found depending on the number of tests performed:

fp_rate = float64(fp) / float64(rounds) * float64(100)

[willf]

Fixed in d9cd4ac