OJ Board

After trying a lot I got it. So many things come to my mind but not the relationship between P(k, m) and P(1, m). When I finally understood it was really simple.

Thanks sclo.

Can anyone help me? I know the problem is solved with DP but my probability skills are limited and I can't think a way of solving it properly! The only way I found that solves the problem is (or something very similar, i didn't coded it, just solved the sample input):

P(k, m) = p0^k + W(k,1)*P(1, m ...

"Input is terminated by a line where the value of n=0. This line should not be processed."

I didn't read the problem very well the first time and got some WA because of this. The sample input doesn't have a 0 at the end :evil:

Input:

1
2000000000
1347838239
123848
137587
38495969
34888888 ...

By different I mean: fabs(r1-r2) > 1e-8. Maybe epsilon could be larger, but that was the only value I tested. And it was enough to make my solution wrong!

But if it isn't a problem in the judge data, then get it as a warning to the ones who are getting WA.

Maybe there's a problem in the judge data. After many WA I got accepted just chaging the way a calculated the radius. Instead of using the first x and y, I used the last ones. I did that after I found out that my program calculated differents radius by the simple formula "radius = sqrt(x^2 + y^2 ...