Conversion Rate for Variant B

The other half of June's checkout traffic at Marlow & Finch was sent to Variant B: the same page, but with the Place order button moved up where it is visible without scrolling and given a solid colour. Visitors were assigned by a coin flip the moment they arrived, so the two groups should differ in nothing except which button they saw.

Same convention as the control arm: the conversion rate here is how often a checkout visit ended in an order, reported as a proportion between 0 and 1 rather than as a percentage. The report is rerun for each month the redesign has been live.

Task: Print Variant B's conversion rate as a single proportion between 0 and 1, unrounded and not scaled into a percentage.

Input

The first line holds one integer, the number of visits to the redesigned page that ended in a completed order. The second line holds one integer, how many people reached the redesigned page in total (at least 1).

Output

One line holding the conversion rate as a proportion, printed the way Python prints a float.

Example:

Input:
65
1000

Output:
0.065