A calibration lab stores long streams of sensor measurements in a compressed ledger to save space: instead of writing every reading, the lab records each distinct reading value once alongside how many times it occurred. Given such a compressed ledger, compute the mean of the full (uncompressed) set of readings, rounded to exactly two decimal places using standard rounding (round half up when the digit immediately after the second decimal place is 5 or more).

Input format

• Line 1: an integer n, the number of ledger rows.
• Each of the next n lines contains two space-separated integers value and count: value occurred count times among the original readings.

Output format

A single line containing the mean of all readings, printed with exactly two digits after the decimal point (for example, 5.00 or 18.33).

Constraints

• 1 <= n <= 1000
• 1 <= value <= 10^6
• 1 <= count <= 10^6
• The same value may appear in more than one row.

3
10 2
20 3
30 1

18.33

2
5 1
5 1

5.00