Problem 47103. Estimate Pi by Monte Carlo method

My solution, although giving the correct answers (checked via "format longg" and accepted by the later tests), did not get accepted in the first test. Maybe there is a problem there?

The test suite has been updated to check against a tolerance; hopefully that will overcome the concern mentioned by Undine Falkenhagen.

Tip: Generate the pair of numbers using p = rand(2,n)' (the order that numbers must be used: rand(n,2) does not work for instance); this will not solve the problem, but it's half way there.