You want a family-wise error rate of alpha across m simultaneous tests. Return the corrected significance threshold each individual test should use.

$$\alpha_{\text{per test}} = \frac{\alpha}{m}$$

Input

alpha = 0.05
m = 20

Output

0.0025