Some new tests were added.
52 authors with inexact calculations got WA.

Edited by author 24.10.2006 10:47

There was an accuracy problem in the old validator. New validator uses long arithmetics. After rejudge some WA submits turned to AC, but authors with real (not integer) numbers in output now have WA.

It is ok that problem is rejudged, but it not ok to delete AC to authors which solution was previously accepted, because is server's fault, and this way we pay the damage.

I think your are wrong.
Admin's tester for 1103 used integer arithm and absolute
correct now and found your mistake. It means that
your solution gives wrong answers in some situations
and, for example, may lead to crach in technical system.
Therefore we must build right program despite of angry.

I think administrators are wrong, because this problem has set of interesting solutions,
and complication of a set of tests they reduce this problems to a writing of long real
arithmetics. i think it is not good. to solve this problem you must solve problems of
accuracy of calculations, instead of thinking up algorithm. Besides it is a senseless
step as calculations in any case are inexact

Edited by author 10.04.2007 10:16

You are wrong. My solution has a lot of float calculations and still I got AC. The calculations should be just accurate enough and you will have no problem.