Anybody who has 1028 accepted, leave a note please. I have
an N*log(N) algorithm that must FLY into the time limit for
N<30000, but it doesn't. It passes all tests from the Ural
championship easily.

> Anybody who has 1028 accepted, leave a note please. I have
> an N*log(N) algorithm that must FLY into the time limit
for
> N<30000, but it doesn't. It passes all tests from the Ural
> championship easily.

IMO self-balancing binary tree would be very good here.

> Anybody who has 1028 accepted, leave a note please. I have
> an N*log(N) algorithm that must FLY into the time limit
for
> N<30000, but it doesn't. It passes all tests from the Ural
> championship easily.

Just use the fact that input is in increasing
order of y and x.

> Anybody who has 1028 accepted, leave a note please. I have
> an N*log(N) algorithm that must FLY into the time limit
for
> N<30000, but it doesn't. It passes all tests from the Ural
> championship easily.

I got O(N^2) really nice and simple algorithm which
really flies like a bird! :). It got 0.1 second
before the time limit :)